Corrections March 2007
UPGRADE
YOUR PHYSICS
NOTES FOR BRITISH SIXTH FORM STUDENTS WHO
 ARE PREPARING FOR THE INTERNATIONAL PHYSICS
OLYMPIAD, OR
 WISH TO TAKE THEIR KNOWLEDGE OF PHYSICS
BEYOND THE A-LEVEL SYLLABI.
A. C. MACHACEK
Page 1
Corrections March 2007
Introduction
The International Physics Olympiad is an annual international physics
competition for pre-university students. Teams of five from each
participating nation attend, and recently over 60 countries have taken
part. Each nation has its own methods for selecting its team members.
In Britain, this is by means of a series of written and practical exams.
The question paper for the first round is circulated to all secondary
schools.
Once the team has been chosen, it is necessary for its members to
broaden their horizons. The syllabus for the International Physics
Olympiad is larger than that of the British A2-level, and indeed forms a
convenient stepping-stone to first year undergraduate work. For this
reason, training is provided to help the British team bridge the gap.
The British Olympiad Committee recognizes the need for teaching
material to help candidates prepare for the international competition.
Furthermore, this material ought to have greater potential in the hands of
students who wish to develop their physics, even if they have no desire
to take part in the examinations.
It is my hope that these notes make a start in providing for this need.
A.C. Machacek, 2001
About the author: Anton Machacek has served on the British Physics
Olympiad Committee since 1997, and has been involved regularly in
training the British team and in writing Physics Challenge examinations.
He served on the academic committee for the International Physics
Olympiad in Leicester in 2000. Anton is Head of Physics at the Royal
Grammar School, High Wycombe, and is an Academic Visitor in the subdepartment of Atomic and Laser Physics, University of Oxford.
Page 2
Corrections March 2007
Contents
1
LINEAR MECHANICS ......................................................................................5
1.1
1.2
1.3
1.4
2
GOING ORBITAL ............................................................................................14
FLUIDS – WHEN THINGS GET STICKY .............................................................17
QUESTIONS ...................................................................................................21
FAST PHYSICS .................................................................................................24
2.1
2.2
2.3
2.4
2.5
3
MOTION IN A LINE ..........................................................................................5
THE PRINCIPLE OF RELATIVITY.....................................................................25
HIGH SPEED OBSERVATIONS.........................................................................25
RELATIVISTIC QUANTITIES ...........................................................................29
THE LORENTZ TRANSFORMS .........................................................................31
QUESTIONS ...................................................................................................35
ROTATION........................................................................................................37
3.1 ANGLE ..........................................................................................................37
3.2 ANGULAR VELOCITY ....................................................................................38
3.3 ANGULAR ACCELERATION ............................................................................38
3.4 TORQUE – ANGULAR FORCE .........................................................................38
3.5 MOMENT OF INERTIA – ANGULAR MASS ......................................................39
3.6 ANGULAR MOMENTUM .................................................................................40
3.7 ANGULAR MOMENTUM OF A SINGLE MASS MOVING IN A STRAIGHT LINE.......41
3.8 ROTATIONAL KINETIC ENERGY ....................................................................42
3.9 SUMMARY OF QUANTITIES............................................................................42
3.10 ROTATIONAL MECHANICS WITH VECTORS .................................................43
3.11 MOTION IN POLAR CO-ORDINATES ...........................................................46
3.12 MOTION OF A RIGID BODY .........................................................................49
3.13 QUESTIONS ...............................................................................................51
4
VIBES, WIGGLES & LIGHT..........................................................................53
4.1
4.2
4.3
4.4
OSCILLATION ................................................................................................53
WAVES & INTERFERENCE .............................................................................55
RAYS.............................................................................................................68
FERMAT’S PRINCIPLE ....................................................................................69
Page 3
Corrections March 2007
4.5
5
QUESTIONS ...................................................................................................69
HOT PHYSICS ..................................................................................................72
5.1 THE CONSERVATION OF ENERGY ..................................................................72
5.2 THE SECOND LAW ........................................................................................73
5.3 HEAT ENGINES AND FRIDGES........................................................................73
5.4 ENTROPY ......................................................................................................76
5.5 IRREVERSIBLE PROCESSES AND THE SECOND LAW .......................................77
5.6 RE-STATEMENT OF FIRST LAW......................................................................78
5.7 THE BOLTZMANN LAW .................................................................................78
5.8 PERFECT GASES ............................................................................................82
5.9 RADIATION OF HEAT .....................................................................................88
5.10 QUESTIONS ...............................................................................................88
6
SPARKS & GENERATION .............................................................................91
6.1
6.2
6.3
6.4
7
CIRCUITS – PUTTING IT TOGETHER ..............................................................109
QUESTIONS .................................................................................................116
WAVES AND PARTICLES ..............................................................................118
UNCERTAINTY ............................................................................................119
ATOMS ........................................................................................................120
LITTLE NUTS ...............................................................................................122
QUESTIONS .................................................................................................124
PRACTICAL PHYSICS .................................................................................126
8.1
8.2
8.3
8.4
8.5
9
MAGNETISM – WHEN THINGS MOVE ..............................................................97
SMALL PHYSICS ...........................................................................................118
7.1
7.2
7.3
7.4
7.5
8
ELECTROSTATICS – WHEN THINGS ARE STILL ................................................91
ERRORS, AND HOW TO MAKE THEM .............................................................126
ERRORS, AND HOW TO MAKE THEM WORSE .................................................128
SYSTEMATIC ERRORS..................................................................................129
WHICH GRAPH? ..........................................................................................130
QUESTIONS .................................................................................................131
APPENDIX.......................................................................................................132
9.1
9.2
MULTIPLYING VECTORS .............................................................................132
DIMENSIONAL ANALYSIS ............................................................................135
Page 4
Corrections March 2007
1 Linear Mechanics
1.1 Motion in a Line
1.1.1
The Fundamentals
1.1.1.1 Kinematics
Mechanics is all about motion. We start with the simplest kind of motion
– the motion of small dots or particles. Such a particle is described
completely by its mass (the amount of stuff it contains) and its position.
There is no internal structure to worry about, and as for rotation, even if it
tried it, no-one would see. The most convenient way of labelling the
position is with a vector r showing its position with respect to some
convenient agreed stationary point.
If the particle is moving, its position will change. If its speed and
direction are steady, then we can write its position after time t as
r = s + ut,
where s is the starting point (the position of the particle at t=0) and u as
the change in position each second – otherwise known as the velocity. If
the velocity is not constant, then we can’t measure it by seeing how far
the object goes in one second, since the velocity will have changed by
then. Rather, we say that u how far the object would go in one second if
the speed or direction remained unchanged that long. In practice, if the
motion remains constant for some small time (called t), and during this
small time, the particle’s position changes r, then the change in position
if this were maintained for a whole second (otherwise known as the
velocity) is
u = r  number of t periods in one second = r  t.
Similarly, if the velocity is changing, we define the acceleration as the
change in velocity each second (if the rate of change of acceleration
were constant. Accordingly, our equation for acceleration becomes
a = u  t.
Hopefully, it is apparent that as the motion becomes more complex, and
the t periods need to be made shorter and shorter, we end up with the
differential equations linking position, velocity and acceleration:
d
r r   u dt
dt
.
d
a  u u   a dt
dt
u
Page 5
Corrections March 2007
1.1.1.2 Dynamics
Now we have a way of describing motion, we need a way of predicting or
explaining the motion which occurs – changing our question from ‘what
is happening?’ to ‘why?’ and our explanation is going to involve the
activity of forces. What do forces do to an object?
The first essential point is that forces are only needed to change (not
maintain) motion. In other words – unless there is a change of velocity,
no force is needed. But how much force is needed?
Newton made the assumption (which we find to be helpful and true) that
the force causes a change in what he called the ‘motion’ –we now call it
momentum. Suppose an object has mass m and velocity u (we shall
clarify what we mean by mass later) – then its momentum is equal to
mu, and is frequently referred to by physicists by the letter p. Newton’s
second law states that if a constant force F is applied to an object for a
short time t, then the change in the momentum is given by F t. In
differential notation d(mu)/dt = F.
In the case of a single object of constant mass it follows that
F
d mu 
du
m
 ma .
dt
dt
His next assumption tells us more about forces and allows us to define
‘mass’ properly. Imagine two bricks are being pulled together by a
strong spring. The brick on the left is being pulled to the right, the brick
on the right is being pulled to the left.
Newton assumed that the force pulling the left brick rightwards is equal
and opposite to the force pulling the right brick leftwards. To use more
mathematical notation, if the force on block no.1 caused by block no.2 is
called f12, then f12=f21. If this were not the case, then if we looked at
the bricks together as a whole object, the two internal forces would not
cancel out, and there would be some ‘left over’ force which could
accelerate the whole object. 1
It makes sense that if the bricks are identical then they will accelerate
together at the same rate. But what if they are not? This is where
Newton’s second law is helpful. If the resultant force on an object of
1
If you want to prove that this is ridiculous, try lifting a large bucket while standing in it.
Page 6
Corrections March 2007
constant mass equals its mass times its acceleration, and if the two
forces are equal and opposite, we say
f12  f 21
m1a1   m2 a 2 ,
a1 m2

a 2 m1
and so the ‘more massive’ block accelerates less. This is the definition
of mass. Using this equation, the mass of any object can be measured
with respect to a standard kilogram. If a mystery mass experiences an
acceleration of 2m/s2 while pushing a standard kilogram in the absence
of other forces, and at the same time the kilogram experiences an
acceleration of 4m/s2 the other way, then the mystery mass must be 2kg.
When we have a group of objects, we have the option of applying
Newton’s law to the objects individually or together. If we take a large
group of objects, we find that the total force
Ftotal   Fi 
i
d
 mi u i
dt i
changes the total momentum (just like the individual forces change the
individual momenta). Note the simplification, though – there are no fij in
the equation. This is because fij + fji = 0, so when we add up the forces,
the internal forces sum to zero, and the total momentum is only affected
by the external forces Fi.
1.1.1.3 Energy and Power
Work is done (or energy is transferred) when a force moves something.
The amount of work done (or amount of energy transformed) is given by
the dot product of the force and the distance moved.
W = F ● r = F r cos 
(1)
where  is the angle between the force vector F and the distance vector
r. This means that if the force is perpendicular to the distance, there is
no work done, no energy is transferred, and no fuel supply is needed.
If the force is constant in time, then equation (1) is all very well and
good, however if the force is changing, we need to break the motion up
into little parts, so that the force is more or less constant for each part.
We may then write, more generally,
W = F  r = F r cos 
Two useful differential equations can be formed from here.
Page 7
(1a)
Corrections March 2007
1.1.1.4 Virtual Work
From equation (1a) it is clear that if the motion is in the direction of the
force applied to the object (i.e. =0), then
W
F,
r
where W is the work done on the object. Accordingly, we can calculate
the force on an object if we know the energy change involved in moving
it. Let’s give an example.
An electron (with charge q) is forced through a resistor (of length L) by a
battery of voltage V. As it goes through, it must lose energy qV, since V
is the energy loss per coulomb of charge passing through the resistor.
Therefore, assuming that the force on the electron is constant (which we
assume by the symmetry of the situation), then the force must be given
by W / d = qV / L. If we define the electric field strength to be the force
per coulomb of charge (F/q), then it follows that the electric field strength
E = V/L.
So far, we have ignored the sign of F. It can not have escaped your
attention that things generally fall downwards – in the direction of
decreasing [gravitational] energy. In equations (1) and (1a), we used the
vector F to represent the externally applied force we use to drag the
object along. In the case of lifting a hodful of bricks to the top of a wall,
this force will be directed upwards. If we are interested in the force of
gravity G acting on the object (whether we drag it or not), this will be in
the opposite direction. Therefore F = G, and
W =  G  r,
G
(1b)
W
.
r
In other words, if an object can lose potential energy by moving from one
place to another, there will always be a force trying to push it in this
direction.
1.1.1.5 Power
Another useful equation can be derived if we differentiate (1a) with
respect to time. The rate of ‘working’ is the power P, and so
P
W F  δr
r

 F .
t
t
t
As we let the time period tend to zero, r/t becomes the velocity, and so
we have:
Page 8
Corrections March 2007
P = F ● u = F u cos 
(2)
where  is now best thought of as the angle between force and direction
of motion. Again we see that if the force is perpendicular to the direction
of motion, no power is needed. This makes sense: think of a bike going
round a corner at constant speed. A force is needed to turn the corner that’s why you lean into the bend, so that a component of your weight
does the job. However no work is done – you don’t need to pedal any
harder, and your speed (and hence kinetic energy) does not change.
Equation (2) is also useful for working out the amount of fuel needed if a
working force is to be maintained. Suppose a car engine is combating a
friction force of 200N, and the car is travelling at a steady 30m/s. The
engine power will be 200N × 30m/s = 6 kW.
Our equation can also be used to derive the kinetic energy. Think of
starting the object from rest, and calculating the work needed to get it
going at speed U. The force, causing the acceleration, will be F=ma.
The work done is given by
W   P dt   F  v dt   m
  mv  dv 

1
2
mv

2 U
0
dv
 v dt
dt
 mU
1
2
(3)
2
although care needs to be taken justifying the integration stage in the
multi-dimensional case. 2
1.1.2
Changing Masses
The application of Newton’s Laws to mechanics problems should pose
you no trouble at all. However there are a couple of extra considerations
which are worth thinking about, and which don’t often get much attention
at school.
The first situation we’ll consider is when the mass of a moving object
changes. In practice the mass of any self-propelling object will change
as it uses up its fuel, and for accurate calculations we need to take this
into account. There are two cases when this must be considered to get
the answer even roughly right – jet aeroplanes and rockets. In the case
of rockets, the fuel probably makes up 90% of the mass, so it must not
be ignored.
2
The proof is interesting. It turns out that v  dv  v dv cos   v dv since
the change in speed dv is equal to |dv| cos where dv (note the bold
type) is the vector giving the change in velocity.
Page 9
Corrections March 2007
Changing mass makes the physics interesting, because you need to
think more carefully about Newton’s second law. There are two ways of
stating it – either
(i)
(ii)
Force on an object is equal to the rate of change of its momentum
Force on an object is equal to mass × acceleration
The first says F  d (mu ) dt  mu  m u  ma  m u , whereas the second
simply states F=ma. Clearly they can’t both be correct, since they are
different. Which is right? The first: which was actually the way Newton
stated it in the first place! The good old F=ma will still work – but you
have to break the rocket into parts (say grams of fuel) – so that the
rocket loses parts, but each part does not lose mass – and then apply
F=ma to each individual part. However if you want to apply a law of
motion to the rocket as a whole, you have to use the more complicated
form of equation.
This may be the first time that you encounter the fact that momentum is
a more ‘friendly’ and fundamental quantity to work with mathematically
than force. We shall see this in a more extreme form when looking at
special relativity.
Let us now try and calculate how a rocket works. We’ll ignore gravity
and resistive forces to start with, and see how fast a rocket will go after it
has burnt some fuel. To work this out we need to know two things – the
exhaust speed of the combustion gas (w), which is always measured
relative to the rocket; and the rate at which the motor burns fuel (in kg/s),
which we shall call .
We’ll think about one part of the motion, when the rocket starts with
mass (M+m), burns mass m of fuel, where m is very small, and in doing
so increases its speed from U to U+u. This is shown below in the
diagram.
Before
M+m
U
After
m
U-w
M
U+u
Notice that the velocity of the burnt fuel is U-w, since w is the speed at
which the combustion gas leaves the rocket (backwards), and we need
to take the rocket speed U into account to find out how fast it is going
relative to the ground.
Conservation of momentum tells us that
Page 10
Corrections March 2007
(M+m) U = m (U-w) + M (U+u)
so
(4)
m w = M u.
We can integrate this expression for u to evaluate the total change in
speed after burning a large amount of fuel. We treat the u (change in U)
as an infinitesimal calculus dU, and the m as a calculus –dM. Notice the
minus sign – clearly the rocket must lose mass as fuel is burnt. Equation
(4) now tells us
w
dM
 dU
M
(5)
This can be integrated to give
1
dM   dU
M
 wln M   U 
 w
M
U final  U initial  w ln initial
M
 final
(6)




This formula (6) is interesting because it tells us that in the absence of
other forces, the gain in rocket speed depends only on the fraction of
rocket mass that is fuel, and the exhaust speed.
In this calculation, we have ignored other forces. This is not a good idea
if we want to work out the motion at blast off, since the Earth’s gravity
plays a major role! In order to take this, or other forces, into account, we
need to calculate the thrust force of the rocket engine – a task we have
avoided so far.
The thrust can be calculated by applying F=ma to the (fixed mass) rocket
M in our original calculation (4). The acceleration is given by dU/t = u/t,
where t is the time taken to burn mass m of fuel. The thrust is
T M
u
mw mw
m
M

 w   w
t
Mt
t
t
(7)
given by the product of the exhaust speed and the rate of burning fuel.
For a rocket of total mass M to take off vertically, T must be greater than
the rocket’s weight Mg. Therefore for lift off to occur at all we must have
w  Mg .
(8)
This explains why ‘heavy’ hydrocarbon fuels are nearly always used for
the first stage of liquid fuel rockets. In the later stages, where absolute
Page 11
Corrections March 2007
thrust is less important, hydrogen is used as it has a better ‘kick per
kilogram’ because of its higher exhaust speed.
1.1.3
Fictitious Forces
Fictitious forces do not exist. So why do we need to give them a
moment’s thought? Well, sometimes they make our life easier. Let’s
have a couple of examples.
1.1.3.1 Centrifugal Force
You may have travelled in one of those fairground rides in which
everyone stands against the inside of the curved wall of a cylinder, which
then rotates about its axis. After a while, the floor drops out – and yet
you don’t fall, because you’re “stuck to the side”. How does this work?
There are two ways of thinking about this. The first is to look at the
situation from the stationary perspective of a friend on the ground. She
sees you rotating, and knows that a centripetal force is needed to keep
you going round – a force pointing towards the centre of the cylinder.
This force is provided by the wall, and pushes you inwards. You feel this
strongly if you’re the rider! And by Newton’s third law it is equally true
that you are pushing outwards on the wall, and this is why you feel like
you are being ‘thrown out’.
While this approach is correct, sometimes it makes the maths easier if
you analyse the situation from the perspective of the rider. Then you
don’t need to worry about the rotation! However in order to get the
working right you have to include an outwards force – to balance the
inward push of the wall. If this were not done, the force from the walls
would throw you into the central space. The outward force is called the
centrifugal force, and is our first example of a fictitious force. It doesn’t
really exist, unless you are working in a rotating reference frame, and
insist that you are at rest.
The difference between the two viewpoints is that in one case the inward
push of the wall provides the centripetal acceleration. In the other it
opposes the centrifugal force - giving zero resultant, and keeping the
rider still. Therefore the formulae used to calculate centripetal force also
give the correct magnitude for centrifugal force. The two differences are:
(i)
Centrifugal force acts outwards, centripetal force acts inwards
(ii) Centrifugal force is only considered if you are assuming that the
cylinder is at rest (in the cylinder’s reference frame). On the other hand,
you only have centripetal accelerations if you do treat the cylinder as a
moving object and work in the reference frame of a stationary observer.
This example also shows that fictitious forces generally act in the
opposite direction to the acceleration that is being ‘ignored’. Here the
Page 12
Corrections March 2007
acceleration is an inward centripetal acceleration, and the fictitious
centrifugal force points outward.
1.1.3.2 Inertial Force
The second example we will look at is the motion of a lift (elevator)
passenger. You know that you ‘feel heavier’ when the lift accelerates
upwards, and ‘feel lighter’ when it accelerates downwards. Therefore if
you want to simplify your maths by treating the lift car as a stationary
box, you must include an extra downward force when the lift is actually
accelerating upwards, and vice-versa. This fictitious force is called the
inertial force. We see again that it acts in the opposite direction to the
acceleration we are trying to ignore.
We shall look more closely at this situation, as it is much clearer
mathematically.
Suppose we want to analyse the motion of a ball, say, thrown in the air
in a lift car while it is accelerating upwards with acceleration A. We use
the vector a to represent the acceleration of the ball as a stationary
observer would measure it, and a’ to represent the acceleration as
measured by someone in the lift. Therefore, a = A + a’. Now this ball
won’t simply travel in a straight line, because forces act on it. Suppose
the force on the ball is F. We want to know what force F’ is needed to
get the right motion if we assume the lift to be at rest.
Newton’s second law tells us that F=ma, if m is the mass of the ball.
Therefore F=m(A+a’), and so F-mA = ma’. Now the force F’ must be
the force needed to give the ball acceleration a’ (the motion relative to
the lift car), and therefore F’=ma’. Combining these equations gives
F’ = F – mA.
(9)
In other words, if working in the reference frame of the lift, you need to
include not only the forces which are really acting on the ball (like
gravity), but also an extra force –mA. This extra force is the inertial
force.
Let us continue this line of thought a little further. Suppose the only
force on the ball was gravity. Therefore F=mg. Notice that
F’ = F – mA = m (g-A)
(10)
and therefore if g=A (that is, the lift is falling like a stone, because some
nasty person has cut the cable), F’=0. In other words, the ball behaves
as if no force (not even gravity) were acting on it, at least from the point
of view of the unfortunate lift passengers. This is why weightlessness is
experienced in free fall.
Page 13
Corrections March 2007
A similar argument can be used to explain the weightlessness of
astronauts in orbiting spacecraft. As stationary observer (or a physics
teacher) would say that there is only one force on the astronauts –
gravity, and that this is just the right size to provide the centripetal force.
The astronaut’s perspective is a little different. He (or she) experiences
two forces – gravity, and the fictitious centrifugal force. These two are
equal and opposite, and as a result they add to zero, and so the
astronaut feels just as weightless as the doomed lift passengers in the
last paragraph.
1.2 Going Orbital
1.2.1
We have the potential
We shall now spend a bit of time reviewing gravity. This is a frequent
topic of Olympiad questions, and is another area in which you should be
able to do well with your A-level knowledge.
Gravitation causes all objects to attract all other objects. To simplify
matters, we start with two small compact masses. The size of the force
of attraction is best described by the equation
Fr  
GMm
R2
(11)
Here G is the Gravitational constant (6.673×10-11 Nm2/kg2), M is the
mass of one object (at the origin of coordinates), and m is the mass of
the other. The equation gives the force experienced by the mass m.
Notice the ‘r’ subscript and the minus sign – the force is radial, and
directed inwards toward the origin (where the mass M is).
It is possible to work out how much work is needed to get the mass m as
far away from M as possible. We use integration




GMm
GMm
 GMm 
R F  dx  R  Fr dr  R r 2 dr   r  R  R .
Notice the use of  Fr in the second stage. In order to separate the
masses we use a force F which acts in opposition to the gravitational
attraction Fr. The equation gives the amount of work done by this force
as it pulls the masses apart.
We usually define the zero of potential energy to be when the masses
have nothing to do with each other (because they are so far away).
Accordingly, the potential energy of the masses m and M is given by
E ( R)  
GMm
.
R
Page 14
(12)
Corrections March 2007
That is, GMm R joules below zero energy. Notice that
Fr ( R)  
dE ( R)
.
dR
(13)
This is a consequence of the definition of work as W   F  dx , and is
generally true. It is useful because it tells us that a forces always point in
the direction of decreasing energy.
The potential energy depends on the mass of both objects as well as the
position. The gravitational potential V(R) is defined as the energy per
unit mass of the second object, and is given by
V R   lim
m0
E ( R, m)
GM
.

m
R
(14)
Accordingly, the potential is a function only of position. The zero limit on
the mass m is needed (in theory) to prevent the small mass disturbing
the field. In practice this will not happen if the masses are fixed in
position. To see the consequences of breaking this rule, think about
measuring the Earth’s gravitational field close to the Moon. If we do this
by measuring the force experienced by a 1kg mass, we will be fine. If
we do it by measuring the force experienced by a 1028kg planet put in
place for the job, we will radically change the motion of Earth and Moon,
and thus affect the measurement.
In a similar way, we evaluate the gravitational field strength as the force
per kilogram of mass. Writing the field strength as g gives
g
MG
R2
(15)
and equation (13) may be rewritten in terms of field and potential as
g ( R)  
1.2.2
dV
.
dR
(16)
Orbital tricks
There is a useful shortcut when doing problems about orbits. Suppose
that an object of mass m is orbiting the centre of co-ordinates, and
experiences an attractive force Fr   Ar n , where A is some constant.
Therefore n=-2 for gravity, and we would have n=+1 for motion of a
particle attached to a spring (the other end fixed at the origin).
Page 15
Corrections March 2007
If the object is performing circular orbits, the centripetal acceleration will
be u 2 R where R is the radius of the orbit. This is provided by the
attractive force mentioned, and so:
mu 2
  Fr  AR n
R
mu 2 AR n 1

2
2
(17)
Now the potential energy E(R) is such that dE dR   Fr  AR n , so
E ( R) 
AR n 1
n 1
(18)
if we take the usual convention that E(R) is zero when the force is zero.
Combining equations (17) and (18) gives
mu 2 n  1

 E ( R)
2
2
(19)
Kinetic Energy × 2 = Potential Energy × (n+1).
(20)
so that
This tells us that for circular gravitational orbits (where n=2), the
potential energy is twice as large as the kinetic energy, and is negative.
For elliptical orbits, the equation still holds: but now in terms of the
average 3 kinetic and potential energies. Equation (20) will not hold
instantaneously at all times for non-circular orbits.
1.2.3
Kepler’s Laws
The motion of the planets in the Solar system was observed extensively
and accurately during the Renaissance, and Kepler formulated three
“laws” to describe what the astronomers saw. For the Olympiad, you
won’t need to be able to derive these laws from the equations of gravity,
but you will need to know them, and use them (without proof).
1.
3
All planets orbit the Sun in elliptical orbits, with the Sun at one
focus.
By average, we refer to the mean energy in time. In other words, if T is the orbital period,
the average of A is given by
1
T

T
0
A(t )dt .
Page 16
Corrections March 2007
2.
The area traced out by the radius of an orbit in one second is the
same for a planet, whatever stage of its orbit it is in. This is
another way of saying that its angular momentum is constant, and
we shall be looking at this in Chapter 3.
3.
The time period of the orbit is related to the [time mean] average
3
radius of the orbit: T  R 2 . It is not too difficult to show that this
is true for circular orbits, but it is also true for elliptic ones.
1.2.4
Large Masses
In our work so far, we have assumed that all masses are very small in
comparison to the distances between them. However, this is not always
the case, as you will often be working with planets, and they are large!
However there are two useful facts about large spheres and spherical
shells. A spherical shell is a shape, like the skin of a balloon, which is
bounded by two concentric spheres of different radius.
1.
The gravitational field experienced at a point outside a sphere or
spherical shell is the same as if all the mass of the shape were
concentrated at its centre.
2.
A spherical shell has no gravitational effect on an object inside it.
These rules only hold if the sphere or shell is of uniform density (strictly –
if the density has spherical symmetry).
Therefore let us work out the gravitational force experienced by a miner
down a very very very deep hole, who is half way to the centre of the
Earth. We can ignore the mass above him, and therefore only count the
bit below him. This is half the radius of the Earth, and therefore has one
eighth of its mass (assuming the Earth has uniform density – which it
doesn’t). Therefore the M in equation (11) has been reduced by a factor
of eight. Also the miner is twice as close to the centre (R has halved),
and therefore by the inverse-square law, we would expect each kilogram
of Earth to attract him four times as strongly. Combining the factors of
1/8 and 4, we arrive at the conclusion that he experiences a gravitational
field ½ that at the Earth’s surface, that is 4.9 N/kg.
1.3 Fluids – when things get sticky
Questions about fluids are really classical mechanics questions. You
can tackle them without any detailed knowledge of fluid mechanics.
There are a few points you need to remember or learn, and that is what
this section contains. Perfect gases are also fluids, but we will deal with
them in chapter 5 – “Hot Physics”.
1.3.1
Floating and ... the opposite
The most important thing to remember is Archimedes’ Principle, which
states that:
Page 17
Corrections March 2007
When an object is immersed in a fluid (liquid or gas), it will
experience an upwards force equal to the weight of fluid
displaced.
By “weight of fluid displaced” we mean the weight of the fluid that would
have been there if the object was not in position. This upward force
(sometimes called the buoyant upthrust) will be equal to
Force = Weight of fluid displaced
= g × Mass of fluid displaced
= g × Density of fluid × Volume of fluid displaced
(21)
For an object that is completely submerged, the “volume of fluid
displaced” is the volume of the object.
For an object that is only partly submerged (like an iceberg or ship), the
“volume of fluid displaced” is the volume of the object below the
“waterline”.
This allows us to find out what will float, and what will sink. If an object is
completely submerged, it will have two forces acting on it. Its weight,
which pulls downwards, and the buoyant upthrust, which pulls upwards.
Upthrust =  V g
Volume V
Mass M
Fluid
Density 
Object floats if:
V>M
 > M/V
Weight = M g
Therefore, things float if their overall density (total mass / total volume) is
less than the density of the fluid. Notice that the overall density may not
be equal to the actual density of the material. To give an example a ship
is made of metal, but contains air, and is therefore able to float because
its overall density is reduced by the air, and is therefore lower than the
density of water. Puncture the hull, and the air is no longer held in place.
Therefore the density of the ship = the density of the steel, and the ship
sinks.
For an object that is floating on the surface of a fluid (like a ship on the
ocean), the upthrust and weight must be equal – otherwise it would rise
Page 18
Corrections March 2007
or fall. From Archimedes’ principle, the weight of water displaced must
equal the total weight of the object.
There is a “brain-teaser” question like this: A boat is floating in the
middle of a lake, and the amount of water in the lake is fixed. The boat
is carrying a large rock. The rock is lifted out of the boat, and dropped
into the lake. Will the level of water in the lake go up or down?
Answer: Level goes down – while the rock was in the boat (and therefore
floating) its weight of water was being displaced. When it was dropped
into the depths, its volume of water was displaced. Now the density of
rock is higher than that of water, so the water level in the lake was higher
in the first case.
1.3.2
Under Pressure
What is the pressure in a fluid? This must depend on how deep you are,
because the deeper you are, the greater weight of fluid you are
supporting. We can think of the pressure (=Force/Area) as the weight of
a square prism of fluid above a horizontal square metre marked out in
the depths.
Pressure = Weight of fluid over 1m2 square
= g × Density × Volume of fluid over 1m2
= g × Density × Depth × Cross sectional area of fluid (1m2)
Pressure = g × Density × Depth
(22)
Of course, this equation assumes that there is nothing pushing down on
the surface of the liquid! If there is, then this must be added in too.
Therefore pressure 10m under the surface of the sea = atmospheric
pressure + weight of a 10m high column of water.
It is wise to take a bit of caution, though, since pressures are often given
relative to atmospheric pressure (i.e. 2MPa more than atmospheric) –
and you need to keep your wits about you to spot whether pressures are
relative (vacuum = -100 kPa) or absolute (vacuum = 0 Pa).
1.3.3
Continuity
Continuity means conservatism! Some things just don’t change – like
energy, momentum, and amount of stuff. This gives us a useful tool.
Think about the diagram below, which shows water in a 10cm [diameter]
pipe being forced into a 5cm pipe.
Page 19
Corrections March 2007
10 cm
5 cm
Water, like most liquids, doesn’t compress much – so it can’t form
bottlenecks. The rate of water flow (cubic metres per second) in the big
pipe must therefore be equal to the rate of water flow in the little pipe.
You might like to draw an analogy with the current in a series circuit.
The light bulb has greater resistance than the wire but the current in both
is the same, because the one feeds the other.
How can we express this mathematically? Let us assume that the pipe
has a cross sectional area A, and the water is going at speed u m/s.
How much water passes a point in 1 second? Let us put a marker in the
water, which moves along with it. In one second it moves u metres.
Therefore volume of water passing a point = volume of cylinder of length
u and cross sectional area A = u A. Therefore
Flow rate (m3/s) = Speed (m/s) × Cross sectional area (m2). (23)
Now we can go back to our original problem. The flow rate in both wide
and narrow pipes must be the same. So if the larger one has twice the
diameter, it has four times the cross sectional area; and so its water
must be travelling four times more slowly.
1.3.4
Bernoulli’s Equation
Something odd is going on in that pipe. As the water squeezes into the
smaller radius, it speeds up. That means that its kinetic energy is
increasing. Where is it getting the energy from? The answer is that it
can only do so if the pressure in the narrower pipe is lower than in the
wider pipe. That way there is an unbalanced force on the fluid in the
cone-shaped part speeding it up. Let’s follow a cubic metre of water
through the system to work out how far the pressure drops.
The fluid in the larger pipe pushes the fluid in the cone to the right. The
force = pressure  area = PL AL. A cubic metre of fluid occupies length
1/AL in the pipe, where AL is the cross sectional area of the pipe to the
left of the constriction. Accordingly, the work done by the fluid in the
wider pipe on the fluid in the cone in pushing the cubic metre through is
Force × Distance = PL AL × 1/AL = PL. However this cubic metre does
work PR AR × 1/AR = PR in getting out the other side. Thus the net
energy gain of the cubic metre is PL  PR, and this must equal the
change in the cubic metre’s kinetic energy uR2/2  uL2/2.
Page 20
Corrections March 2007
1.3.5
The Flow Equation
Equation (23) is also useful in the context of electric currents, and can be
adapted into the so-called flow equation. Let us suppose that the fluid
contains charged particles. Suppose that there are N of these particles
per cubic metre of fluid, and each particle has a charge of q coulombs,
then:
Current = Flow rate of charge (charge / second)
= Charge per cubic metre (C/m3)  flow rate (m3/s)
= N q  Area  Speed .
(24)
Among other things, this equation shows why the free electrons in a
semiconducting material travel faster than those in a metal. If the
semiconductor is in series with the metal, the current in both must be the
same. However, the free charge density N is much smaller in the
semiconductor, so the speed must be greater to compensate.
1.4 Questions
1.
Calculate the work done in pedalling a bicycle 300m up a road inclined
at 5° to the horizontal.
2.
Calculate the power of engine when a locomotive pulls a train of 200
000kg up a 2° incline at a speed of 30m/s. Ignore the friction in the
bearings. +
3.
A trolley can move up and down a track. It’s potential energy is given by
V = k x4, where x is the distance of the trolley from the centre of the
track. Derive an expression for the force exerted on the trolley at any
point. +
4.
A ball bearing rests on a ramp fixed to the top of a car which is
accelerating horizontally. The position of the ball bearing relative to the
ramp is used as a measure of the acceleration of the car. Show that if
the acceleration is to be proportional to the horizontal distance moved by
the ball (measured relative to the ramp), then the ramp must be curved
upwards in the shape of a parabola. ++
5.
Use arguments similar to equation (3) to prove that the kinetic energy is
still given by 12 mu 2 even when the force which has caused the
acceleration from rest has not been applied uniformly in a constant
direction. +
6.
Calculate the final velocity of a rocket 60% of whose launch mass is
propellant, where the exhaust velocity is 2000m/s.
Repeat the
calculation for a rocket where the propellant makes up 90% of the launch
mass. In both cases neglect gravity.
Page 21
Corrections March 2007
7.
Repeat question 6, now assuming that rockets need to move vertically in
a uniform gravitational field of 9.8N/kg. Calculate the velocity at MECO
(main engine cut-off) and the greatest height reached. Assume that both
rockets have a mass of 10 000kg on the launch pad, and that the
propellant is consumed evenly over one minute. ++
8.
A 70kg woman stands on a set of bathroom scales in an elevator.
Calculate the reading on the scales when the elevator starts accelerating
upwards at 2m/s2, when the elevator is going up at a steady speed, and
when the elevator decelerates at 2m/s2 before coming to a halt at the top
floor of the building.
9.
The woman in q8 is a juggler. Describe how she might have to adjust
her throwing techniques in the elevator as it accelerates and
decelerates.
10. Architectural models can not be properly tested for strength because
they appear to be stronger than the real thing. To see why, consider a
half-scale model of a building made out of the same materials. The
weight is 1/8 of the real building, but the columns are ¼ the cross
sectional area. Accordingly the stress on the columns is half of that in
the full size building, and accordingly the model can withstand much
more severe load before collapsing. To correct for this, a 1:300
architectural model is put on the end of a centrifuge arm of radius 10m
which is spun around.
The spinning ‘simulates’ an increased
gravitational force which allows the model to be accurately tested. How
many times will the centrifuge go round each minute?
11. Consider an incompressible fluid flowing from a 15cm diameter pipe into
a 5cm diameter pipe. If the velocity and pressure before the constriction
are 1m/s and 10 000 N/m2, calculate the velocity and pressure in the
constricted pipe. Neglect the effects of viscosity and turbulent flow. To
work out the new pressure, remember that the increase in speed
involves an increase of kinetic energy, and this energy must come from
somewhere – so there will be a drop in pressure.
12. Calculate the orbital time period T of a satellite skimming the surface of a
planet with radius R and made of a material with density . Calculate
the orbital speed for an astronaut skimming the surface of a comet with a
10km radius.
13. The alcohol percentage in wine can be determined from its density. A
very light glass test tube (of cross sectional area 0.5cm2) has 5g of lead
pellets fixed to the bottom. You place the tube in the wine, lead first, and
it floats with the open end of the tube above the surface of the wine.
You can read the % alcohol from markings on the side of the tube.
Calculate how far above the lead the 0% and 100% marks should be
placed. The density of water is 1.00g/cm3, while that of ethanol is
Page 22
Corrections March 2007
1.98g/cm3. Where should the 50% line go? Remember that alcohol
percentages are always volume percentages. +
Page 23
Corrections March 2007
2 Fast Physics
Imagine a summer’s day. You are sunbathing by the side of a busy
motorway while you wait for a pickup truck to rescue your car, which has
broken down. All of a sudden, an irresponsible person throws a used
drinks can out of their car window, and it heads in your direction. To
make things worse, they were speeding at the time. Ouch.
The faster the car was going, the more it will hurt when the can hits you.
This is because the can automatically takes up the speed the car was
travelling at. Suppose the irresponsible person could throw the can at
10mph, and their car is going at 80mph. The speed of the can, as you
see it, is 90mph if it was thrown forwards, and 70mph if it was thrown
backwards.
To sum this up,
Velocity as measured by you = Velocity of car + Velocity of throwing
where we use velocities rather than speeds so that the directionality can
be taken into account.
So far, this probably seems very obvious. However, let’s extend the
logic a bit further. Rather than a car, let us have a star, and in place of
the drinks can, a beam of light. Many stars travel towards us at high
speeds, and emit light as they do so. We can measure the speed of this
light in a laboratory on Earth, and compare it with the speed of ‘ordinary’
light made in a stationary light bulb. And the worrying thing is that the
two speeds are the same.
No matter how hard we try to change it, light always goes at the same
speed. 4 This tells us that although our ideas of adding velocities are
nice and straightforward, they are also wrong. In short, there is a
problem with the Newtonian picture of motion. This problem is most
obvious in the case of light, but it also occurs when anything else starts
travelling very quickly.
While this is not the way Einstein approached the problem, it is our way
into one of his early theories – the Special Theory of Relativity – and it is
part of the Olympiad syllabus.
Before we go further and talk about what does happen when things go
fast, please be aware of one thing. These observations will seem very
4
Light does travel different speeds in different materials. However if the measurement is
made in the same material (say, air or vacuum) the speed registered will always be the same,
no matter what we do with the source.
Page 24
Corrections March 2007
weird if you haven’t read them before. But don’t dismiss relativity as
nonsense just because it seems weird – it is a better description of
Nature than classical mechanics – and as such it demands our respect
and attention.
2.1 The Principle of Relativity
The theory of special relativity, like all theories, is founded on a premise
or axiom. This axiom cannot be ‘derived’ – it is a guessed statement,
which is the starting point for the maths and the philosophy. In the case
of special relativity, the axiom must be helpful because its logical
consequences agree well with experiments.
This principle, or axiom, can be stated in several ways, but they are
effectively the same.
1.
There is no method for measuring absolute (non-relative) velocity.
The absolute speed of a car cannot be measured by any method at
all. On the other hand, the speed of the car relative to a speed
gun, the Earth, or the Sun can all be determined.
2.
Since it can’t be measured – there is no such thing as absolute
velocity.
3.
The ‘laws of physics’ hold in all non-accelerating laboratories 5 ,
however ‘fast’ they may be going. This follows from statement 2,
since if experiments only worked for one particular laboratory
speed, that would somehow be a special speed, and absolute
velocities could be determined relative to it.
4.
Maxwell’s theory of electromagnetism, which predicts the speed of
light, counts as a law of physics. Therefore all laboratories will
agree on the speed of light. It doesn’t matter where or how the
light was made, nor how fast the laboratory is moving.
2.2 High Speed Observations
In this section we are going to state what relativity predicts, as far as it
affects simple observations. Please note that we are not deriving these
statements from the principles in the last section, although this can be
done. For the moment just try and understand what the statements
mean. That is a hard enough job. Once you can use them, we shall
then worry about where they come from.
5
We say non-accelerating for a good reason. If the laboratory were accelerating, you would
feel the ‘inertial force’, and thus you would be able to measure this acceleration, and indeed
adjust the laboratory’s motion until it were zero. However there is no equivalent way of
measuring absolute speed.
Page 25
Corrections March 2007
2.2.1
Speeding objects look shortened in the direction of
motion.
A metre stick comes hurtling towards you at high speed. With a clever
arrangement of cameras and timers, you are able to measure its length
as it passes you. If the stick’s length is perpendicular to the direction of
travel, you still measure the length as 1 metre.
However, if the stick is parallel to its motion, it will seem shorter to you.
If we call the stick’s actual length (as the stick sees it) as L0, and the
apparent length (as you measure it) La, we find
L a  L0 1 
u2
,
c2
(1)
where u is the speed of the metre stick relative to the observer. The
object in the square root appears frequently in relativistic work, and to
make our equations more concise, we write

1
1  u c 
2
(2)
so that equation (1) appears in shorter form as
La 
2.2.2
L0

.
(3)
Speeding clocks tick slowly
A second observation is that if a clock whizzes past you, and you use
another clever arrangement of timers and cameras to watch it, it will
appear (to you) to be going slowly.
We may state this mathematically. Let T0 be a time interval as
measured by our (stationary) clock, and let Ta be the time interval as we
see it measured by the whizzing clock.
Ta 
2.2.3
T0

(4)
Slowing and shrinking go together
Equations (3) and (4) are consistent – you can’t have one without the
other. To see why this is the case, let us suppose that Andrew and Betty
both have excellent clocks and metre sticks, and they wish to measure
their relative speed as they pass each other. They must agree on the
relative speed. Andrew times how long it takes Betty to travel along his
metre stick, and Betty does the same.
Page 26
Corrections March 2007
The question is: how does Andrew settle his mind about Betty’s
calculation? As far as he is concerned, she has a short metre stick, and
a slow clock – how can she possibly get the answer right! Very easily –
providing that her clock runs ‘slow’ by the same amount that her metre
stick is ‘short’ 6 .
An experimental example may help clarify this. Muons are charged
particles that are not stable, and decay with a half-life of 2s. Because
they are charged, you can accelerate them to high speeds using a large
electric field in a particle accelerator. You can then measure how far
they travel down a tube before decaying. Given that ‘the laws of physics
are the same in all reference frames’, this must mean that muon and
experimenter agree on the position in the tube at which the muon passes
away.
The muon gets much further down the tube than a classical calculation
would predict, however the reason for this can be explained in two ways:

According to the experimenter, the muon is travelling fast, so it has
a slow clock, and therefore lives longer – so it can get further.

According to the muon, it still has a woefully short life, but the tube
(which is whizzing past) is shorter so it appears to get further along
in the 2s.
For the two calculations to agree, the ‘clock slowing’ must be at the
same rate as the ‘tube shrinkage’.
2.2.4
Speeding adds weight to the argument
The most useful observation of them all, as far as the Olympiad syllabus
is concerned is this: if someone throws a 1kg bag of sugar at you at high
speed, and you (somehow) manage to measure its mass as it passes,
you will register more than 1kg.
If the actual mass of the object is M0, and the apparent mass is Ma, we
find that
Ma  M0 .
(5)
The actual mass is usually called the ‘rest mass’ – in other words the
mass as measured by an observer who is at rest with respect to the
object.
6
Note that ‘slow’ and ‘short’ are placed in quotation marks. Betty’s clock and metre stick are
not defective – however to Andrew they appear to be.
Page 27
Corrections March 2007
2.2.4.1 The Universal Speed Limit
This formula has important consequences. First of all, this is the origin
of the ‘universal speed limit’, which is a well-known consequence of
special relativity. This states that you will never measure the speed of
an object (relative to you) as being greater than the speed of light.
Let us pause for a moment to see why. Suppose the object concerned
is an electron in a particle accelerator (electrons currently hold the speed
record on Earth for the fastest humanly accelerated objects). It starts at
rest with a mass of about 10-30 kg. We turn on a large, constant electric
field, and the electron starts to move relative to the accelerator.
However, as it gets close to the speed of light, it starts to appear more
massive. Therefore since our electric field (hence accelerating force) is
constant, the electron’s acceleration decreases. In fact, the acceleration
tends to zero as time passes, although it never reaches zero exactly
after a finite time. We are never able to persuade the electron to break
the ‘light-barrier’, since when u  c ,    , and the apparent mass
becomes very large (so the object becomes impossible to accelerate any
further).
Please note that this does not mean that faster-than-light speeds can
never be obtained. If we accelerate one electron to 0.6c Eastwards, and
another to 0.6c Westwards, the approach speed of the two electrons is
clearly superlumic (1.2c) as we measure it with Earth-bound
speedometers. However, even in this case we find that the velocity of
one of the electrons as measured by the other is still less than the speed
of light. This is a consequence of our first observation – namely that
relative velocities do not add in a simple way when the objects are
moving quickly.
In fact the approach speed, as the electrons see it, is 0.882c. If you
want to perform these calculations, the formula turns out to be
u AC 
u AB  u BC
1  u AB u BC c 2
,
(6)
where uAB means the velocity of B as measured by A. Equation (6) only
applies when all three relative velocities are parallel (or antiparallel).
2.2.4.2 Newton’s Law of motion
Our second consequence is that we need to take great care when using
Newton’s laws. We need to remember that the correct form of the
second law is
F
d
momentum
dt
Page 28
(7)
Corrections March 2007
Why is care needed? Look closely for the trap – if the object speeds up,
its mass will increase. Therefore the time derivative of the mass needs
to be included as well as the time derivative of the velocity. We shall
postpone further discussion until we have had a better look at
momentum.
2.3 Relativistic Quantities
Now that we have mentioned the business of relativistic mass increase,
it is time to address the relativistic forms of other quantities.
2.3.1
Momentum
Momentum is conserved in relativistic collisions, providing we define it as
the product of the apparent mass and the velocity.
p   m0 u
(7)
Notice that when you use momentum conservation in collisions, you will
have to watch the  factors. Since these are functions of the speed u,
they will change if the speed changes.
2.3.2
Force
The force on a particle is the time derivative of its momentum. Therefore
F
d 
d
 d
p  m0   u  u
.
dt
dt 
 dt
(8)
In the case where the speed is not changing,  will stay constant, and
the equation reduces to the much more straightforward F=m0a. One
example is the motion of an electron in a magnetic field.
2.3.3
Kinetic Energy
Now that we have an expression for force, we can integrate it with
respect to distance to obtain the work done in accelerating a particle. As
shown in section 1.1.1, this will give the kinetic energy of the particle.
We obtain the result 7
7
If you wish to derive this yourself, here are the stages you need. Firstly, differentiate  with
respect to u to convince yourself that
 3u
d
u


du c 2 1  u 2 c 2 3 2
c2

du c 2

.
d  3 u
Using this result, the derivation can be completed (see over the page):
Page 29
Corrections March 2007
K    1m 0 c 2 .
(9)
This states that the gain in energy of a particle when accelerated is
equal to the gain in mass × c2. From this we postulate that any increase
in energy is accompanied by a change in mass. The argument works
backwards too. When stationary, the particle had mass m0. Surely
therefore, it had energy m0c2 when at rest.
We therefore write the total energy of a particle as
E  K  m0 c 2   m0 c 2 .
2.3.4
(10)
A Relativistic Toolkit
We can derive a very useful relationship from (10), (7) and the definition
of :

E 2  p 2 c 2   2 m02 c 2 c 2  v 2

  v 2 
  m c 1     .
 c 


2 4
 m0 c
2
2
0
(11)
4
This is useful, since it relates E and p without involving the nasty  factor.
Another equation which has no gammas in it can be derived by dividing
momentum by total energy:
p  mo u
u

 2,
2
E  m0 c
c
(12)
which is useful if you know the momentum and total energy, and wish to
know the speed.
2.3.5
Tackling problems
If you have to solve a ‘collision’ type problem, avoid using speeds at all
costs. If you insist on having speeds in your equations, you will also
have gammas, and therefore headaches. So use the momenta and
energies of the individual particles in your equations instead. Put more
bluntly, you should write lots of ‘p’s, and ‘E’s, but no ‘u’s. Use the
d
du
dt   m 0 u
dt
dt
dt


du 
c2 
c2
d  m 0  u  u  2 d  m 0  u d   m0 c 2
 m 0  u  u  
d 
u
 u


K   Fdx   Fu dt   m 0 u 2

Page 30

Corrections March 2007
conservation laws to help you. In relativistic work, you can always use
the conservation of E – even in non-elastic collisions. The interesting
thing is that in an inelastic collision, you will find the rest masses greater
after the collision.
To obtain the values you want, you need an equation which relates E
and p, and this is provided by (11). Notice in particular that the quantity
 E 2   p 2 c 2 when applied to a group of particles has two things to
commend it.

Firstly, it is only a function of total energy and momentum, and
therefore will remain the same before and after the collision.

Secondly, it is a function of the rest masses (see equation 11) and
therefore will be the same in all reference frames.
Finally, if the question asks you for the final speeds, use (12) to calculate
them from the momenta and energies.
2.4 The Lorentz Transforms
The facts outlined above (without the derivations) will give you all the
information you need to tackle International Olympiad problems.
However, you may be interested to find out how the observations of
section 2.2 follow from the general assumptions of section 2.1. A full
justification would require a whole book on relativity, however we can
give a brief introduction to the method here.
We start by stating a general problem. Consider two frames of reference
(or co-ordinate systems) – Andrew’s perspective (t,x,y,z), and Betty’s
perspective (t’,x’,y’,z’). We assume that Betty is shooting past Andrew in
the +x direction at speed v. Suppose an ‘event’ happens, and Andrew
measures its co-ordinates. How do we work out the co-ordinates Betty
will measure?
The relationship between the two sets of co-ordinates is called the
Lorentz transformation, and this can be derived as shown below:
2.4.1
Derivation of the Lorentz Transformation
We begin with the assumption that the co-ordinate transforms must be
linear. The reason for this can be illustrated by considering length,
although a similar argument works for time as well. Suppose that
Andrew has two measuring sticks joined end to end, one of length L1
and one of length L2. He wants to work out how long Betty reckons they
are. Suppose the transformation function is T. Therefore Betty
measures the first rod as T(L1) and the second as T(L2). She therefore
will see that the total length of the rods is T(L1) + T(L2). This must also
be equal to T(L1+L2), since L1+L2 is the length of the whole rod
Page 31
Corrections March 2007
according to Andrew. Since T(L1+L2) = T(L1) +T(L2), the transformation
function is linear.
We can now get to work. Let us consider Betty’s frame of reference to
be moving in the +x direction at speed v, as measured by Andrew. Betty
will therefore see Andrew moving in her –x direction at the same speed.
To distinguish Betty’s co-ordinates from Andrew’s, we give hers dashes.
Given the linear nature of the transformation, we write
 x   A B  x 
   
 
 t    C D  t 
where A, B, C and D are functions of the relative velocity +v (i.e. Betty’s
velocity as measured by Andrew).
There must also be an inverse transformation
 x 1  D
   
 t  d  C
 B  x  
 
A  t  
where d is the determinant of the first matrix.
Now this second matrix is in itself a transformation for a relative velocity
–v, and therefore should be of a very similar form to the first matrix. We
find that the only way we can ensure that there is symmetry between the
two is to make the determinant equal to one (d=1). We shall therefore
assume this from here on.
Next we consider what happens if x’=0. In other words we are tracing
out Betty’s motion as Andrew sees it. Therefore we must have x=vt.
Using the first matrix, this tells us that B=-vA. A similar argument on the
second matrix – where we must have x’=-vt’ where Betty now watches
Andrew’s motion [x=0], gives –Dv = B = -vA. Therefore A=D.
We now have B and D expressed in terms of A, so the next job is to
work out what C is. This can be done since we know that the
determinant AD – BC = 1. Therefore we find that
C
1  A2
.
vA
Page 32
Corrections March 2007
Summarizing, our matrix is now expressed totally in terms of the
unknown variable A. We may calculate it by remembering that both
Andrew and Betty will agree on the speed of travel, c, of a ray of light.
Andrew will express this as x=ct, Betty would say x’=ct’, but both must
be valid ways of describing the motion. Therefore
 ct    A B  ct 
 
   
 t    C D  t 
ct  Ac  B
c

t  Cc  D
Cc  D c  Ac  B
Cc 2  B since A  D
1 - A2  2
c  vA

 vA 
1  A 2 c 2  v 2 A 2
1
A
v2
1 2
c


This concludes our reasoning, and gives the Lorentz transforms (after a
little algebra to evaluate C) as:
x     x  vt 
xv 

t   t  2 
 c 
x    x   vt 
.
x v 

t   t  2 
c 

1

 v  2 

1 
  c  


We have not considered any other dimensions here, however the
transformation here is easy since Andrew and Betty agree on all lengths
in the y and z directions. In other words y’=y, z’=z. This is a necessary
consequence of the principle of relativity: the distance between the ends
of a rod held perpendicular to the direction of motion can be measured
simultaneously in all frames of reference. If this agreed measurement
was different to that of an identical rod in a different frame, the observers
would be able to work out which of them was ‘moving’ and which of them
was still.
Page 33
Corrections March 2007
2.4.2
Using the Lorentz Transforms
Having these transforms at our disposal, we can now derive the
‘shrinking rod’ and ‘slowing clock’ equations.
Suppose Betty is holding a stick (of length L) parallel to the x-axis. We
want to know how long Andrew thinks it is. To measure it, he will
measure where the ends of the rod are at a particular moment, and will
then measure the distance between these points. Clearly the two
positions need to be measured simultaneously in his frame of reference,
and thus t is the same for both measurements. We know from that Betty
thinks it has length L, and therefore x’=L. Using the first of the Lorentz
equations (the one which links x’, x and t), and remembering that t is the
same for both measurements,
x   x
Lapparent 
L.

Similarly we may show how a clock appears to slow down. Betty is
carrying the clock, so it is stationary with respect to her, and x’ (her
measurement of the clock’s position) will therefore be constant. The
time interval shown on Betty’s clock is t’, while Andrew’s own clock will
measure time t. Here t’ is the time Andrew sees elapsing on Betty’s
clock, and as such is equal to Tapparent. Using the fourth Lorentz equation
(the one with x’, t and t’ in it), and remembering that x’ remains constant,
we have
t  t 
Tapparent 
2.4.3
T.

Four Vectors
The Lorentz transforms show you how to work out the relationships
between the (t,x,y,z) co-ordinates measured in different frames of
reference. We describe anything that transforms in the same way as a
four vector, although strictly speaking we use (ct,x,y,z) so that all the
components of the vector have the same units. Three other examples of
four vectors are:

(c, ux, uy, uz) is called the four velocity of an object, and is the
derivative of (ct, x,y,z) with respect to the proper time . Proper
time is the time elapsed as measured in the rest frame of the object
t=.

(mc,px,py,pz) the momentum four vector. Here m is equal to m0.
This must be a four vector since it is equal to the rest mass
Page 34
Corrections March 2007
multiplied by the four velocity (which we already know to be a four
vector).

(/c,kx,ky,kz) the wave four vector, where  is the angular
frequency of the wave (=2f), and k is a vector which points in the
direction the wave is going, and has magnitude 2/. This can be
derived from the momentum four vector in the case of a photon,
since the momentum and total energy of a photon are related by
E=pc, and the quantum theory states that E=hf=h/2 and
p=h/=hk/2.
It also turns out that the dot product of any two four-vectors is ‘frameinvariant’ – in other words all observers will agree on its value. The dot
product of two four-vectors is slightly different to the conventional dot
product, as shown below:
ct , x, y, z   ct , x, y, z   x 2  y 2  z 2  ct 2 .
Notice that we subtract the product of the first elements.
The dot product of the position four vector with the wave four vector
gives
ct , x, y, z   
c , k x , k y , k z   k  r  t .
Now this is the phase of the wave, and since all observers must agree
whether a particular point is a peak, a trough or somewhere in between,
then the phase must be an invariant quantity. Accordingly, since
(ct,x,y,z) makes this invariant when ‘dotted’ with (/c,kx,ky,kz), it follows
that (/c,kx,ky,kz) must be a four vector too.
2.5 Questions
1. Work out the relativistic  factor for speeds of 1%, 50%, 90% and 99% of
the speed of light.
2. Work out the speeds needed to give  factors of 1.0, 1.1, 2.0, 10.0.
3. A muon travels at 90% of the speed of light down a pipe in a particle
accelerator at a steady speed. How far would you expect it to travel in 2s
(a) without taking relativity into account, and (b) taking relativity into
account?
4. A particle with rest mass m and momentum p collides with a stationary
particle of mass M. The result is a single new particle of rest mass R.
Calculate R in terms of p and M.
Page 35
Corrections March 2007
5. The principal runway at the spaceport on Arcturus-3 has white squares of
side length 10m painted on it. A set of light sensors on the base of a
spacecraft can take a ‘picture’ of the whole runway at the same time.
What will the squares look like in the image if the spacecraft is passing the
runway at a very high speed? Each sensor takes a picture of the runway
directly underneath it, so you do not need to take into account the different
times taken by light to reach the sensors from different parts of the
runway.
6. When an electron is accelerated through a voltage V, its kinetic energy is
given by eV where e is the size of the charge on the electron and is equal
to 1.61019C. Taking the mass of the electron to be 9.11031kg, work out
(a) the kinetic energy and speed of the electron when V=511kV (b) the
kinetic energy and speed when V=20kV (c) the percentage error in the
kinetic energy for V=20kV when calculated using the non-relativistic
equation ½ mu2.
7. Prove that the kinetic energy of a particle of rest mass m and speed u is
given by ½ mu2 if the speed is small enough in comparison to the speed of
light. Work out the speed at which the non-relativistic calculation would be
in error by 1%.
8. Suppose a spacecraft accelerates with constant acceleration a (as
measured by the spacecraft’s onboard accelerometers). At t=0 it is at rest
with respect to a planet. Work out its speed relative to the planet as a
function of time (a) as measured by clocks on the spacecraft, and (b) as
measured by clocks on the planet. Note that the instantaneous speed of
the craft relative to the planet will be agreed upon by spacecraft and
planet.
Page 36
Corrections March 2007
3 Rotation
Rotational motion is all around us [groan] – from the acts of subatomic
particles, to the motion of galaxies. Calculations involving rotations are
no harder than linear mechanics; however the quantities we shall be
talking about will be unfamiliar at first. Having already studied linear
mechanics, you will be at a tremendous advantage, since we shall find
that each ‘rule’ in linear mechanics has its rotational equivalent.
3.1 Angle
In linear mechanics, the most fundamental measurement is the position
of the particle. The equivalent base of all rotational analysis is angle: the
question “How far has the car moved?” being exchanged for “How far
has the wheel gone round?” – a question which can only be answered
by giving an angle. In mechanics, the radian is used for measuring
angles. While you may be more familiar with the degree, the radian has
many advantages.
We shall start, then by defining what we mean by a radian. Consider a
sector of a circle, as in the diagram; and let the circle have a radius r.
The length of the arc, that is the curved line in the sector, is clearly
related to the angle. If the angle were made twice as large, the arc
length would also double.
Can we use arc length to measure the angle? Not as it stands, since we
haven’t taken into account the radius of the circle. Even for a fixed angle
(say 30°), the arc will be longer on a larger circle. We therefore define
the angle (in radians) as the arc length divided by the
Arc length = r if  is
circle radius. Alternatively you might say that the angle
measured in rad ians
in radians is equal to the length of the arc of a unit circle
(that is a circle of 1m radius) that is cut by the angle.
r
Notice one simplification that this brings. If a wheel, of
radius R, rolls a distance d along a road, the angle the
wheel has turned through is given by d/R in radians.
Were you to calculate the angle in degrees, there would
be nasty factors of 180 and  in the answer.

Before getting too involved with radians, however, we must work out a
conversion factor so that angles in degrees can be expressed in radians.
To do this, remember that a full circle (360°) has a circumference or arc
length of 2r. So 360°=2 rad. Therefore, 1 radian is equivalent to
(360/2)° = (180/)°.
Page 37
Corrections March 2007
3.2 Angular Velocity
Having discussed angle as the rotational equivalent of position, we now
turn our attention to speed. In linear work, speeds are given in metres
per second – the distance moved in unit time. For rotation, we speak of
‘angular velocity’, which tells us how fast something is spinning: how
many radians it turns through in one second. The angular velocity can
also be thought of as the derivative of angle with respect to time, and as
such is sometimes written as  , however more
commonly the Greek letter  is used, and the dot
v sin 
is avoided. To check your understanding of this,
v
try and show that 1 rpm (revolution per minute) is

equivalent to /30 rad/s, while one cycle per
second is equivalent to 2 rad/s.
r
Now remember the definition of angle in radians,
and that the distance moved by a point on the rim
of a wheel will move a distance s = r when the wheel rotates by an
angle . The speed of the point will therefore be given by u = ds/dt = r
d/dt = r.
For a point that is not fixed to the wheel, the situation is a little more
complex. Suppose that the point has a velocity v, which makes an angle
 to the radius (as in the figure above). We then separate v into two
components, one radial (v cos ) and one rotational (v sin ). Clearly the
latter is the only one that contributes to the angular velocity, and
therefore in this more general case, v sin  = r.
3.3 Angular acceleration
It should come as no surprise that the angular acceleration is the time
derivative of , and represents the change in angular velocity (in rad/s)
divided by the time taken for the change (in s). It is measured in rad/s2,
and denoted by  or  or  . For an object fastened to the rim of a
wheel, the ‘actual’ acceleration round the rim (a) will be given by a=du/dt
= r d/dt = r, while for an object not fastened, we have a sin  = r. 8
3.4 Torque – Angular Force
Before we can start ‘doing mechanics’ with angles, we need to consider
the rotational equivalent of force – the amount of twist. Often a twist can
be applied to a system by a linear force, and this gives us a ‘way in’ to
the analysis. We say that the strength of the twist is called the ‘moment’
of the force, and is equal to the size of the force multiplied by the
distance from pivot to the point where the force acts. A complication
8
Here we are not including the centripetal acceleration which is directed towards the centre of
the rotation.
Page 38
Corrections March 2007
F
arises if the force is not tangential – clearly a force acting
along the radius of a wheel will not turn it – and so our
simple ‘moment’ equation needs modifying. 9
There are two ways of proceeding, and they yield the
same answer. Suppose the force F makes an angle 
with the radius. We can break this down into two
r sin 
components – one of magnitude F cos , which is radial
and does no turning; and the other, tangential
component (which does contribute to the turning) of
magnitude F sin . The moment or torque only includes the relevant
component, and so the torque is given by C = Fr sin .
r

The alternative way of viewing the situation is not to measure the
distance from the centre to the point at which the force is applied.
Instead, we draw the force as a long line, and to take the distance as the
perpendicular distance from force line to centre. The diagram shows
that this new distance is given by r sin , and since the force here is
completely tangential, we may write the moment or torque as the product
of the full force and this perpendicular distance – i.e. C = F r sin , as
before.
3.5 Moment of Inertia – Angular Mass
Of the three base quantities of motion, namely distance, mass and time,
only time may be used with impunity in rotational problems. We now
have an angular equivalent for distance (namely angle), so the next task
is to determine an angular equivalent for mass.
This can be done by analogy with linear mechanics, where the mass of
an object in kilograms can be determined by pushing an object, and
calculating the ratio of the applied force to the acceleration it caused: m
= F/a. Given that we now have angular equivalents for force and
acceleration, we can use these to find out the ‘angular mass’.
Think about a ball of mass m fixed to the rim of a wheel that is
accelerating with angular acceleration . We
shall ignore the mass of the wheel itself for now.
F sin  Now let us push the mass round the wheel with a
F
force F. Therefore we calculate the ‘angular

mass’ I by
9
Why force × radius? We can use a virtual work argument (as in section 1.1.1.4) to help us.
Suppose a tangential force F is applied at radius r. When the object moves round by angle ,
it moves a distance d = r, and the work done by the force = Fd = Fr = Fr × angular force
× angular distance. Now since energy must be the same sort of thing with rotational motion
as linear, the rotational equivalent of force must be Fr.
Page 39
Corrections March 2007
I
C


rF sin 
F
 r2  r2 m
a sin  r 
a
where we have used the fact that the mass m will be the ratio of the
force F to the linear acceleration a, as dictated by Newton’s Second
Law. This formula can also be used for solid objects, however in this
case, the radius r will be the perpendicular distance from the mass to the
axis. The total ‘angular mass’ of the object is calculated by adding up
the I = m r2 from each of the points it is made from.
Usually this ‘angular mass’ is called the moment of inertia of the object.
Notice that it doesn’t just depend on the mass, but also on the distance
from the point to the centre. Therefore the moment of inertia of an object
depends on the axis it is spun round.
An object may have a high angular inertia, therefore, for two reasons.
Either it is heavy in its own right; or for a lighter object, the mass is a
long way from the axis.
3.6 Angular Momentum
In linear motion, we make frequent use of the ‘momentum’ of objects.
The momentum is given by mass × velocity, and changes when a force
is applied to the object. The force applied, is in fact the time derivative of
the momentum (provided that the mass doesn’t change). Frequent use
is made of the fact that total momentum is conserved in collisions,
provided that there is no external force acting.
It would be useful to find a similar ‘thing’ for angular motion. The most
sensible starting guess is to try ‘angular mass’ × angular velocity. We
shall call this the angular momentum, and give it the symbol L = I . Let
us now investigate how the angular momentum changes when a torque
is applied. For the moment, assume that I remains constant.
d
dL d
 I  I
 I  C
dt
dt dt
Thus we see that, like in linear motion, the time derivative of angular
momentum is ‘angular force’ or torque. Two of the important facts that
stem from this statement are:
1.
If there is no torque C, the angular momentum will not change.
Notice that radial forces have C = 0, and therefore will not change the
angular momentum. This result may seem unimportant – but think of the
planets in their orbits round the Sun. The tremendous force exerted on
them by the Sun’s gravity is radial, and therefore does not change their
angular momenta even a smidgen. We can therefore calculate the
velocity of planets at different parts of their orbits using the fact that the
angular momentum will remain the same. This principle also holds when
Page 40
Corrections March 2007
scientists calculate the path of space probes sent out to investigate the
Solar System.
2.
The calculation above assumes that the moment of inertia I of the
object remains the same. This seems sensible, after all, in a linear
collision, the instantaneous change of a single object’s mass would be
bizarre 10 , and therefore we don’t need to guard against the possibility of
a change in mass when we write F = dp/dt.
In the case of angular motion, this situation is different. The
moment of inertia can be changed, simply by rearranging the mass of
the object closer to the axis. Clearly there is no external torque in doing
this, so we should expect the angular momentum to stay the same. But
if the mass has been moved closer to the axis, I will have got smaller.
Therefore  must have got bigger. The object will now be spinning
faster! This is what happens when a spinning ice dancer brings in
her/his arms – and the corresponding increase in revs. per minute is well
known to ice enthusiasts and TV viewers alike.
To take an example, suppose that all the masses were moved twice as
close to the axis. The value of r would halve, so I would be quartered.
We should therefore expect  to get four times larger. This is in fact
what happens.
3.7 Angular momentum of a single mass moving in a straight
line
If we wished to calculate the angular momentum of a planet in its orbit
round the Sun, we need to know how L is related to the linear speed v.
This is what we will now work out.
Using the same ideas as in figure 2, the velocity v will have both radial
and ‘rotational’ components. The rotational component will be equal to v
sin , while the radial component cannot contribute to the angular
momentum. It is the rotational component that corresponds to the speed
of a mass fixed to the rim of a wheel, and as such is equal to radius ×
angular velocity. Thus v sin  = r . So the angular momentum
L  I  mr 2 
v sin 
 mvr sin 
r
10
Two cautions. Firstly, in a rocket, the mass of the rocket does decrease as the burnt fuel is
chucked out the back, however the total mass does not change. Therefore F=dp/dt=ma still
works, we just need to be careful that the force F acts on (and only on) the stuff included in
the mass m. A complication does arise when objects start travelling at a good fraction of the
speed of light – but this is dealt with in the section on Special Relativity.
Page 41
Corrections March 2007
is given by the product of the mass, the radius and the rotational (or
tangential) component of the velocity.
For an object on a straight line path, this can also be stated (using figure
3) as the mass × speed × distance of closest approach to centre.
3.8 Rotational Kinetic Energy
Lastly, we come to the calculation of the rotational kinetic energy. We
may calculate this by adding up the linear kinetic energies of the parts of
the object as the spin round the axis. Notice that in this calculation, as
the objects are purely rotating, we shall assume  = /2 – i.e. there is no
radial motion.
K  12 mv 2  12 mr   12 mr 2 2  12 I 2
2
We see that the kinetic energy is given by half the angular mass ×
angular velocity squared – which is a direct equivalent with the half mass
× speed2 of linear motion.
3.9 Summary of Quantities
Quantity
Symbol
Unit
Definition
Other equations
Angular velocity

rad/s
 = d/dt
r  = v sin 
Angular
acceleration

rad/s2
 = d/dt
r  = a sin 
Torque
C
Nm
C = F r sin 
Moment of inertia
I
kg m2
I=C/
I = m r2
Angular
momentum
L
kg m2 /s
L=I
L = m v r sin 
Rot. Kinetic
Energy
K
J
K = I 2 / 2
K = ½m (v sin )2
Page 42
Corrections March 2007
3.10
Rotational mechanics with vectors
This section involves much more advanced mathematics, and you will be
able to get by in Olympiad problems perfectly well without it. However, if
you like vectors and matrices, read on...
So far we have just considered rotations in one plane – that of the paper.
In general, of course, rotations can occur about any axis, and to describe
this three dimensional situation, we use vectors. With velocity v,
momentum p and force F, there is an obvious direction – the direction of
motion, or the direction of the ‘push’. With rotation, the ‘direction’ is less
clear.
Imagine a clock face on this paper, with the minute hand rotating
clockwise. What direction do we associate with this motion? Up
towards 12 o’clock because the hand sometimes points that way?
Towards 3 o’clock because the hand sometimes points that way? Both
are equally ridiculous. In fact the only way of choosing a direction that
will always apply is to assign the rotation ‘direction’ perpendicular to the
clock face – the direction in which the hands never point.
This has not resolved our difficulty completely. Should the arrow point
upwards out of the paper, or down into it? After thought we realise that
one should be used for clockwise and one for anti-clockwise motion, but
which way? There is no way of proceeding based on logic – we just
have to accept a convention. The custom is to say that for a clockwise
rotation, the ‘direction’ is down away from us, and for anticlockwise
rotation, the direction is up towards us.
Various aides-memoire have been presented for this – my favourite is to
consider a screw. When turned clockwise it moves away from you:
when turned anticlockwise it moves towards you. For this reason the
convention is sometimes called the
‘right hand screw rule’.
r
r sin 
With this convention established, we

can now use vectors for angular
velocity , angular momentum L, and
torque C. Kinetic energy, like in linear
motion, is a scalar and therefore needs
no further attention. The moment of
inertia I is more complex, and we shall
come to that later.
Let us consider the angular velocity first. If we already know  and r,
what is v, assuming that only rotational velocities are allowed?
Remembering that w must point along the axis of the rotation, we may
draw the diagram above, which shows that the radius of the circle that
Page 43
Corrections March 2007
our particle actually traces out is r sin  where  is the angle between r
and . 11 This factor of sin  did not arise before in this way, since our
motion was restricted to the plane which contained the centre point, and
thus  = /2 for all our 2-dimensional work. Therefore the velocity is
equal to w multiplied by the radius of the circle traced out, i.e. v =  r sin
. This may be put on a solid mathematical foundation using the vector
cross product namely v = ×r. This is our first vector identity for
rotational motion.
By a similar method, we may analyse the acceleration. We come to the
corresponding conclusion a =  × r. 12
Next we tackle torque. Noting our direction convention, and our earlier
equation C = F r sin , we set C = r × F. Similarly, from L = (mv) r sin  =
p r sin , we set L = r × p.
With these three vector equations we may get to work. Firstly, notice:
d
d
d
d
L  r  p   v  p  r  p  0  r  mv   r  ma  r  F  C
dt
dt
dt
dt
The time derivative of angular momentum is the torque, as before.
Notice too that the (v×p) term disappears since p has the same direction
as v, and the vector cross product of two parallel vectors is zero.
3.10.1.1 General Moment of Inertia
Our next task is to work out the moment of inertia. This can be more
complex, since it is not a vector. Previously we defined I by the
relationships C = I , and also used the expression L = I . Now that C,
, L and  are vectors, we conclude that I must be a matrix, since a
vector is made when I is multiplied by the vectors  or . Our aim is to
find the matrix that does the job.
For this, we use our vector equations v =  × r and C = r × F, we let the
components of r be (x,y,z), and we also use the mathematical result that
for any three vectors A, B and C, A  B  C  A  CB  A  B  C .
11
We use  to represent the angle between r and , to distinguish it from the angle  between
r and v, which is of course a right angle for a strict rotation.
12
This intentionally does not include the centripetal acceleration, as before. If you aim to
calculate this a from the former equation v =  × r, then you get a = dv/dt = d(×r)/dt = ×r +
×v = ×r + ×(×r) = ×r +  (r.) – r 2. The final two terms in this equation deal with the
centripetal acceleration. However in real situations, the centripetal force is usually provided
by internal or reaction forces, so often problems are simplified by not including it.
Page 44
Corrections March 2007
C  r  F  mr  a 
 mr  α  r 
 mr 2 α  mr  α r
r 2  x2

 m  xy
 
xz

 xy
r  y2
 yz
 xz 

 yz  α

r2  z2 
 y2  z2

 m  xy
  xz

 xy
x z
 yz
 xz 

 yz  α
x 2  y 2 
2
2
2
This result looks horrible. However let us simplify matters by aligning
our axes so that the z axis is the axis of the acceleration . In other
words  = (0,0,). We now have
  xz 


C  m  yz  α
 x2  y2 


which is a little better. Notice that it is still pretty nasty in that the torque
required to cause this z-rotation acceleration is not necessarily in the zdirection! Another consequence of this is that the angular momentum L
is not necessarily parallel to the angular velocity . However for many
objects, we rotate them about an axis of symmetry. In this case the xz
and yz terms become zero when summed for all the masses in the
object, and what we are left with is the mass multiplied by the distance
from the axis to the masses (that is x2 + y2). Alternatively, for a flat
object (called a lamina) which has no thickness in the z direction, the xz
and yz terms are zero anyway, because z=0.
At this point, you are perfectly justified in saying ‘yuk’ and sticking to twodimensional problems. However this result we have just looked at has
interesting consequences. When a 3-d object has little symmetry, it can
roll around in some very odd ways. Some of the asteroids and planetary
moons in our Solar System are cases in point.
The moment of inertia can also be obtained from the rotational
momentum, however, the form is identical to that worked out above from
Newton’s second law, as shown here.
L  r  p  mr  v 
 mr  ω  r 
 mr 2 ω  mr  ω r
The calculation then proceeds as before.
Page 45
Corrections March 2007
3.10.1.2 General Kinetic Energy
Our final detail is kinetic energy. This can be calculated using v =  × r,
and the vector rule that A  B  C  B  C  Α  .
K  12 mv 2  12 mv  v 
 12 mv  ω  r 
 12 mω  r  v 
 12 ω  r  mv 
 12 ω  r  p 
 12 ω  L  12 ω  Iω
For the cases where I can be simplified, this reduces to the familiar form
K = I 2/2.
3.11
Motion in Polar Co-ordinates
When a system is rotating, it often makes sense to use polar coordinates. In other words, we characterise position by its distance from
the centre of rotation (i.e. the radius r) and by the angle  it has turned
through.
Conversion between these co-ordinates and our usual
Cartesian (x,y) form are given by simple trigonometry:
x  r cos
y  r sin 
(1)
y
θ̂
r̂
r

x
When analysing motion problems, though, there are complications if
polar co-ordinates are used. These stem from the fact that the
‘increasing r’ and ‘increasing ’ directions themselves depend on the
value of , as we shall see. Let us start by defining the vector r to be the
position of a particle relative to some convenient origin. The length of
this vector r gives the distance from particle to origin. We define r̂ to be
a unit vector parallel to r. Similarly, we define the vector θˆ to be a unit
vector pointing in the direction the particle would have to go in order to
increase  while keeping r constant. Let us now evaluate the time
derivative of r – in other words, let’s find the velocity of the particle:
Page 46
Corrections March 2007
d r rˆ  dr
d rˆ
d rˆ
d
r
 rˆ  r
 r rˆ  r
,
dt
dt
dt
dt
dt
(2)
where we have used the dot above a letter to mean ‘time derivative of’.
Now if the particle does not change its , then the direction r̂ will not
change either, and we have a velocity given simply by r r̂ . We next
consider the case when r doesn’t change, and the particle goes in a
circle around the origin. In this case, our formula would say that the
d r̂
. We know from section 3.2 that in this case, the
velocity was r
dt
speed is given by r, that is r , so the velocity will be r θ̂ . In order to
make this agree with our equation for dr/dt, we would need to say that
d
rˆ   θˆ .
dt
(3)
Does this make sense? If you think about it for a moment, you should
find that it does. Look at the diagram below. Here the angle  has
changed a small amount . The old and new rˆ vectors are shown, and
form two sides of an isosceles triangle, the angle between them being
. Given that the sides rˆ have length 1, the length of the third side is
going to be approximately equal to  (with the approximation getting
better the smaller  is). Notice also that the third side – the vector
corresponding to rˆnew  rˆold is pointing in the direction of θˆ . This allows
us to justify statement (1).
r
r̂new
r̂old

θ̂
In a similar way, we may show that
d ˆ
θ   rˆ .
dt
Remembering that our velocity is given by
v  r  r rˆ  r θˆ ,
we may calculate the acceleration as
Page 47
(4)
Corrections March 2007
d rˆ
d θˆ
a  v  rrˆ  r θˆ  rθˆ  r
 r
dt
dt
2
 rrˆ  r θˆ  rθˆ  r  θˆ  r rˆ
.
 r  r 2 rˆ  r  2r θˆ

 
(5)

Now suppose that a force acting on the particle (with mass m), had a
radial component Fr, and a tangential component F. We could then
write



Fr  m r  r 2
.
F  m r  2r

(6)

There are many consequences of these equations for rotational motion.
Here are three:
1.
For an object to go round in a circle (that is r staying constant, so that
r  r  0 ), we require a non-zero radial force Fr  mr 2 . The minus
sign indicates that the force is to be in the opposite direction to r, in
other words pointing towards the centre. This, of course, is the
centripetal force needed to keep an object going around in a circle at
constant speed.
2.
If the force is purely radial (we call this a central force), like gravitational
attraction, then F= 0. It follows that
0  mr  2mr
 mr 2  2mrr ,


d
mr 2 
dt
(7)

and accordingly the angular momentum mr 2  mr 2 does not
change. This ought to be no surprise, since we found in section 3.6
that angular momenta are only changed if there is a torque, and a
radial force has zero torque.
3.
One consequence of the conservation of angular momentum is the
apparently odd behaviour of an object coming obliquely towards the
centre (that is, it gets closer to the origin, but is not aimed to hit it).
Since r decreases,  must increase, and this is what happens – in fact
the square term causes  to quadruple when r halves.

We can analyse this in terms of forces using (6): F  m r  2r

when F= 0. Since r is decreasing, while  increases, the non-zero
value of the 2mr term gives rise to a non-zero  , and hence an
acceleration of rotation. If you were sitting next to the particle at the
Page 48
Corrections March 2007
time, you would wonder what caused it to speed up, and you would
think that there must have been a force acting upon it.
This is another example of a fictitious force (see section 1.1.3), and is
called the Coriolis force. It is used, among other things, to explain why
the air rushing in to fill a low pressure area of the atmosphere begins to
rotate – thus setting up a ‘cyclone’. Some people have attempted to
use the equation to explain the direction of rotation of the whirlpool you
get above the plughole in a bath.
Put very bluntly – the Coriolis force is the force needed to ‘keep’ the
object going in a true straight line. Of course, a stationary observer
would see no force – after all things go in straight lines when there are
no sideways forces acting on them. The perspective of a rotating
observer is not as clear – and this Coriolis force will be felt to be as real
as the centrifugal force discussed in section 1.1.3.1.
3.12
Motion of a rigid body
When you are dealing with a rigid body, things are simplified in that it
can only do two things – move in a line and rotate. If forces Fi are
applied to positions ri on a solid object free to move, its motion is
completely described by

a linear acceleration given by a   Fi M , where M is the total
mass of the body, and


a rotational acceleration given by α   ri  Fi I about a point
called the centre of mass, where ri’ is the position of point i
relative to the centre of mass and I is the moment of inertia of
the object about the axis of rotation. 13
This means, among other things, that the centre of mass itself moves as
if it were a point particle of mass M. In turn, if a force is applied to the
object at the centre of mass, it will cause the body to move with a linear
acceleration, without any rotational acceleration at all.
The proof goes as follows. Suppose the object is made up of lots of
points ri (of mass mi) fixed together. It follows that Newton’s second law
states (as in section 1.1.1.2)
13
This assumes that the angular acceleration is a simple speeding up or slowing down of an
existing rotation. If  and  are not parallel, the situation is more complex.
Page 49
Corrections March 2007
d ( mi u i )
  Fi
dt
d 2 (mi ri )
 dt 2   Fi
d 2  mi ri
  Fi
dt 2

Now suppose we define the position R such that MR =  mi ri, then it
follows that
M
d 2R
 Ftotal
dt 2
and the point R moves as if it were a single point of mass M being acted
on by the total force. This position R is called the centre of mass.
Given that we already know that R does not have any rotational motion,
this must be the centre of rotation, and we can use the equation from
section 3.10 to show that the rate of change of angular momentum of the
object about this point, d(I)/dt, is equal to the total torque  (ri – R)×Fi
acting on the body about the point R. Given that the masses don’t
change, we may write
d
mi u i  Fi   f ij
dt
j
mi a i  Fi   f ij
.
j
mi ri  a i  ri  Fi  ri   f ij
j
 m r  a  r  F  r  f
i i
i
i
i
i
i
i
ij
ij
The final term sums to zero since fij+fji=0, and the internal forces
between two particles must either constitute a repulsion, an attraction or
the two forces must occur at the same place. In any of these cases fij ×
(ri–rj) = 0.
If we now express the positions ri in terms of the centre of mass position
R and a relative position ri’, where ri = R + ri’ (so ai = A+ai’), then
Page 50
Corrections March 2007
 


   A  a i     R  ri   Fi

 






 mi R  a i   mi ri  A   mi ri  a i   R  Fi   ri  Fi
 m  R  r
i
Rm a
i
i
i



 0   mi ri  a i   R  Fi   ri  Fi
m r

i i


 a i   ri  Fi
since miri’ = mi(ri–R) = MR – MR = 0. Now, as shown earlier,
d mi ri  u i  dt  mi u i  u i  mi ri  a i  0  mi ri  a i , and so
d


ri  u i   ri  Fi

dt
d
d

L   Iω   ri  Fi  C
dt
dt
and so the rate of change of angular momentum about the centre of
mass is given by the total moment of the external forces about the centre
of mass.
3.13
Questions
1. A car has wheels with radius 30cm. The car travels 42km. By what angle
have the wheels rotated during the journey? Make sure that you give your
answer in radians and in degrees.
2. Why does the gravitational attraction to the Sun not change the angular
momentum of the Earth?
3. Calculate the speed of a satellite orbiting the Earth at a distance of 42
000km from the Earth’s centre.
4. A space agency plans to build a spacecraft in the form of a cylinder 50m in
radius. The cylinder will be spun so that astronauts inside can walk on the
inside of the curved surface as if in a gravitational field of 9.8 N/kg.
Calculate the angular velocity needed to achieve this.
5. A television company wants to put a satellite into a 42 000km radius orbit
round the Earth. The satellite is launched into a circular low-Earth orbit
200km above the Earth’s surface, and a rocket motor then speeds it up. It
then coasts until it is in the 42000km orbit with the correct speed. How
fast does it need to be going in the low-Earth orbit in order to coast up to
the correct position and speed?
6. Estimate the gain in angular velocity when an ice-skater draws her hands
in towards her body.
Page 51
Corrections March 2007
7. One theory of planet formation says that the Earth was once a liquid
globule which gradually solidified, and its rotation as a liquid caused it to
bulge outwards in the middle – a situation which remains to this day: the
equatorial radius of the Earth is about 20km larger than the polar radius. If
the theory were correct, what would the rotation rate of the Earth have
been just before the crust solidified? Assume that the liquid globule was
sufficiently viscous that it was all rotating at the same angular velocity.
Page 52
Corrections March 2007
4 Vibes, Wiggles & Light
4.1 Oscillation
Any system in stable equilibrium can be persuaded to oscillate. If it is
removed from the equilibrium, there will be a force (or other influence)
that attempts to maintain the status quo. The size of the force will
depend on the amount of the disturbance.
Suppose that the disturbance is called x. The restoring force can be
written


F   Ax  Bx 2  Cx 3  ,
(1)
where the minus sign indicates that the force acts in the opposite
direction to the disturbance. If x is small enough, x2 and x3 will be so
small that they can be neglected. We then have a restoring force
proportional to the displacement x.
Just because the system has a force acting to restore the equilibrium,
this does not mean that it will return to x=0 immediately. All systems
have some inertia, or reluctance to act quickly. For a literal force, this
inertia is the mass of the system – and we know that the acceleration
caused by a force (F) is given by F/m, where m is the mass. We can
therefore work on equation (1) to find out more:
d 2x
F  m 2   Ax
dt
.
2
d x
A
 x
m
dt 2
(2)
This differential equation has the solution:
x  x 0 cost   
 Am
,
(3)
which is indeed an oscillation. We are using x0 to denote the amplitude.
Notice that, as we are working in radians, the cosine function needs to
advance 2 to go through a whole cycle. Therefore we can work out the
time period (T) and frequency (f):
T  2
2
T
.

1 
f  
T 2
Page 53
(4)
Corrections March 2007
Seeing that =2f, we notice that  is none other than the angular
frequency of the oscillation, as defined in chapter 3.
These equations are perfectly general, and so whenever you come
across a system with a differential equation like (2), you know the
system will oscillate, and furthermore you can calculate the frequency.
4.1.1
Non-linearity
Equation (1) has left an unanswered question. What happens if x is big
enough that x2 and x3 can’t be neglected? Clearly solution (3) will no
longer work. In fact the equation probably won’t have a simple solution,
and the system will start doing some really outrageous things. Given
that it has quadratic terms in it, we say it is non-linear; and a non-linear
equation will send most physics students running away, screaming for
mercy.
Let me give you an example. There are very nice materials that look
harmlessly transparent. However they are designed so that the nonlinear terms are very important when light passes through them. The
result – you put red laser light in, and it comes out blue (at twice the
frequency). They are called ‘doubling crystals’ and are the sort of thing
that might freak out an unsuspecting GCSE examiner.
Our world would be much less wonderful if it were purely linear – no
swirls in smoke, no wave-breaking (and hence surfing), and extremely
boring weather – not to mention rigid population dynamics. While the
non-linear terms add to the spice of life, I for one am grateful that many
phenomena can be well described using linear equations. Otherwise
physics would be much more frustrating, and bridge design would be just
as hard as predicting the weather.
4.1.2
Energy
Before we move from oscillations to waves, let us make one further
observation. The energy involved in the oscillation is proportional to the
square of the amplitude. We shall show this in two ways.
First: If the displacement is given by equation (3), we notice that the
velocity is given by
u  x   x 0 sin t    .
(5)
At the moment when the system passes through its equilibrium (x=0)
point, all the energy is in kinetic form. Therefore the total energy is
E  K  x  x 0   12 mu 2  12 m 2 x 02
which is indeed proportional to the amplitude squared.
Page 54
(6)
Corrections March 2007
Second: When the displacement is at its maximum, there is no kinetic
energy. The energy will all be in potential form. We can work out the
potential energy in the system at displacement x, by evaluating the work
done to get it there:
E pot   Fdx   Axdx 

1
2

Ax 2 .
(7)
Notice that we did not include the minus sign on the force. This is
because when we work out the ‘work done’ the force involved is the
force of us pulling the system. This is equal and opposite to the
restoring force of the system, and as such is positive (directed in the
same direction as x).
The total energy is given by the potential energy at the moment when x
has its maximum (i.e. x=x0). Therefore
E  E pot  x  x 0   12 Ax 02 .
(8)
Equations (8) and (6) are in agreement. This can be shown by inserting
the value of  from equation (3) into (6).
While we have only demonstrated that energy is proportional to
amplitude squared for an oscillation, it turns out that the same is true for
linked oscillators – and hence for waves. The intensity of a wave (joules
of energy transmitted per second) is proportional to the amplitude
squared in exactly the same way.
Intensity of a wave is also related to another wave property – its speed.
The intensity is equal to the amount of energy stored on a length u of
wave, where u is the speed. This is because this is the energy that will
pass a point in one second (a length u of wave will pass in this time).
4.2 Waves & Interference
The most wonderful property of waves is that they can interfere. You
can add three and four and get six, or one, or 4.567, depending on the
phase relationship between the two waves. You can visualize this using
either trigonometry or vectors (phasors). However, before we look at
interference in detail, we analyse a general wave.
4.2.1
Wave number
Firstly, we define a useful parameter called the wave number. This is
usually given the letter k, and is defined as
k
2

,
Page 55
(13)
Corrections March 2007
where  is the wavelength. If we write the shape of a ‘paused’ wave as
y=A cos(), the phase  of a wave is given by
  kD .
(14)
We can see that this makes sense by combining equations (13) and
(14):
  kD 
2D

.
(15)
If the distance D is equal to a whole wavelength, we expect the wave to
be doing the same thing as it was at =0. And since cos(2)=cos(0), this
is indeed the case.
A variation on the theme is possible. You may also see wave vectors k:
these have magnitude as defined in (13), and point in the direction of
energy transfer.
4.2.2
Wave equations
We are now in a position to write a general equation for the motion of a
wave with angular frequency  and wave number k:
y  A cost  kx    .
We can check that this is correct, since

if we look at a particular point (value of x), and watch as time passes,
we will pass from one peak to the next when t =2f t has got bigger
by 2 (i.e. t=1/f as it should).

if we look at a particular moment in time (value of t), and look at the
position of adjacent peaks, they should be separated by one
wavelength = 2/k. Now for adjacent peaks, the values of kx will
differ by 2according to the formula above, and so this is correct.

if we follow a particular peak on the wave – say the one where t–
kx+f=0, we notice that x=(t+f)/k = t/k + constant, and hence the
position moves to increasing x at a speed equal to /k, as indeed it
should since /k = 2f/(2/) = f = v.
It follows that a leftwards-travelling wave has a function which looks like
y  A cost  kx    .
Page 56
Corrections March 2007
4.2.3
Standing waves
Imagine we have two waves of equal amplitude passing along a string in
the two different directions. The total effect of both waves is given by
adding them up:
y  A cost  kx  1   A cost  kx   2 
 2 cost  12 1   2   coskx  12  2  1 
At any time, the peaks and troughs will only occur at the places where
the second cosine is +1 or –1, and so the positions of the peaks and
troughs do not change. This is why this kind of situation is called a
standing wave. While there is motion, described by the first cosine, the
positions of constructive interference between the two counterpropagating waves remain fixed (these are called antinodes), as to the
positions of destructive interference (the nodes).
While there are many situations which involve counter-propagating
waves, this usually is caused by the reflection of waves at boundaries
(like the ends of a guitar string). Accordingly, there is nothing keeping
the phase constants 1 and 2 the same, and so the standing wave
doesn’t develop. However if the frequency is just right, then it works, as
indicated in section 4.2.7.5.
4.2.4
Trigonometric Interference
We are now in a position to look at the fundamental property of waves –
namely interference. Our first method of analysis uses trigonometry.
Suppose two waves arrive at the same point, and are described by
x1  A cost  and x 2  B cost    respectively.
To find out the
resulting sum, we add the two disturbances together.
X  x1  x 2
 A cos t  B cost   
 A cos t  B cos t cos   B sin t sin 
  A  B cos   cos t  B sin  sin t
 X 0 cos  cos t  sin  sin t 
(9)
 X 0 cos 0 t   
where we define
X0 

cos  
 A  B cos  2  B sin  2
A 2  B 2  2 AB cos 
A  B cos 
X0
sin  
Page 57
B sin 
X0
.
(10)
Corrections March 2007
The amplitude of the resultant is given by X0. Notice that if A=B, the
expression simplifies:
X 0  A 2  2 cos 
 A 2 1  cos  
 A 4 cos 2  12  
,
(11)
 2 A cos 12  
and we obtain the familiar result that if the waves are ‘in phase’ (=0),
the amplitude doubles, and if the waves are  radians (half a cycle) ‘out
of phase’, we have complete destructive interference.
Equation (10) can be used to provide a more general form of this
statement – the minimum resultant amplitude possible is |A-B|, while the
maximum amplitude possible is A+B.
This statement is reminiscent of the ‘triangle inequality’, where the length
of one side of a triangle is limited by a similar constraint on the lengths of
the other two sides. This brings us to our second method of working out
interferences: by a graphical method.
4.2.5
Graphic Interference
In the graphic method a vector represents each wave. The length of the
vector gives the amplitude, and the relative orientation of two vectors
indicates their phase relationship. If the phase relationship is zero, the
two vectors are parallel, and the total length is equal to the sum of the
individual lengths. If the two waves are  out of phase, the vectors will
be antiparallel, and so will partly (or if A=B, completely) cancel each
other out.
The diagram below shows the addition of two waves, as in the situation
above. Notice that since + = , cos  = –cos. One application of
the cosine rule gives
X0 
A 2  B 2  2 AB cos 
in agreement with equation (10).
X
B


A
Page 58
(12)
Corrections March 2007
4.2.6
Summary of Interference Principles
The results of the last section allow us to determine the amplitude once
we know the phase difference between the two waves. Usually the two
waves have come from a common source, but have travelled different
distances to reach the point. Let us suppose that the difference in
distances is D – this is sometimes called the path difference. What will 
be?
To find out, we use the wave number k. The phase difference  is given
by
  kD 
2D

.
(14)
If the distance D is equal to a whole wavelength, we expect the two
waves to interfere constructively, since peak will meet peak, and trough
will meet trough. In equation (15), if D=, then =2, and constructive
interference is indeed obtained, as can be seen from equation (12).
Similarly, we find that if D is equal to /2, then =, and equation (12)
gives destructive interference.
4.2.7
Instances of two-wave interference
4.2.7.1 Young’s “Two Slit Experiment”
Two cases need to be dealt with. The first is known as the two-slit
experiment, and concerns two sources in phase, which are a distance d
apart, as shown in the diagram below. The path difference is given by
D  d sin  , in the case that d is much smaller than the distance from
sources to observer. Using the conditions in the last section, we see
that interference will be constructive if D  d sin   n where n is an
integer.
To obser vation
point

d

D = d sin 
4.2.7.2 Thin films and colours on soap bubbles
The second case is known as thin film interference, and concerns the
situation in the diagram below. Here the light can take one of two
routes.
Page 59
Corrections March 2007

B
t
A
C
t

E
The path difference is calculated:
D  AC  CB
 AC  CE
 2t cos
Before we can work out the conditions required for constructive or
destructive interference, there is an extra caution to be borne in mind –
the reflections.
4.2.7.3 Hard & Soft Reflections
The reflection of a wave from a surface (or more accurately, the
boundary between two materials) can be hard or soft.

Hard reflections occur when, at the boundary, the wave passes into
a ‘sterner’ material. At these reflections, a peak (before the
reflection) becomes a trough (afterwards) and vice-versa. This is
usually stated as “a  phase difference is added to the wave by the
reflection.” These mean the same thing since cos      cos .
To visualize this – imagine that you are holding one end of a rope,
and a friend sends a wave down the rope towards you. You keep
your hand still. At your hand, the incoming and outgoing waves
interfere, but must sum to zero (after all, your hand is not moving,
so neither can the end of the rope). Therefore if the incoming wave
is above the rope, the outgoing wave must be below. In this way,
peak becomes trough and vice-versa.

Soft reflections, on the other hand, are where the boundary is from
the ‘sterner’ material. At these reflections, a peak remains a peak,
and there is no phase difference to be added.
Page 60
Corrections March 2007
What do we mean by ‘sterner’? Technically, this is a measurement of
the restoring forces in the oscillations which link to produce the wave –
the A coefficients of (1). However, the following table will help you to get
a feel for ‘sternness’.
Wave
From
To
Reflection
Light
Reflection off mirror
Light
Air
Water / Glass
Hard
Light
Water / Glass
Air
Soft
Light
Lower
index
Sound
Solid / liquid
Air
Soft
Sound
Air
Solid / liquid
Hard
Wave on string
Reflection off fastened end of string
Hard
Wave on string
Reflection off unsecured end of string
Soft
Hard
refractive Higher refractive Hard
index
4.2.7.4 Film Interference Revisited
Going back to our thin film interference: sometimes both reflections will
be hard; sometimes one will be hard, and the other soft.
The formulae for constructive interference are:
Both reflections hard, or both soft:
One reflection hard, one soft:
D  2t cos  n
D  2t cos  12   n
(16)
(17)
The difference comes about because of the phase change on reflection
at a hard boundary.
4.2.7.5 Standing Waves
Equations (16) and (17) with =0 can be used to work out the
wavelengths allowed for standing waves. For a standing wave, we must
have constructive interference between a wave and itself (having
bounced once back and forth along the length of the device). The
conditions for constructive interference in a pipe, or on a string of length
L (round trip total path = 2L) are
Page 61
Corrections March 2007
2 L  n soft reflections at both ends
2 L  n  12  soft reflection at one end
2 L  n  1 hard reflections at both ends.
4.2.7.6 Two waves, different frequencies
All the instances given so far have involved two waves of identical
frequencies (and hence constant phase difference). What if the
frequencies are different? Let us suppose that our two waves are
described by x1  A cos 1t and x 2  B cos  2 t , where we shall write
   2  1 . When we add them, we get:
X  x1  x 2
 A cos 1t  B cos t cos 1t  B sin t sin 1t
  A  B cos t  cos 1t  B sin t sin 1t
,
(18)
 X 0 cos1t   
where
X0 

cos  
 A  B cos t 2  B sin t 2
A 2  B 2  2 AB cos t
.
(19)
A  B cos t
X0
We see that the effective amplitude fluctuates, with angular frequency .
On the other hand, if the two original waves had very different
frequencies, then this fluctuation may be too quick to be picked up by the
detector. In this case, the resultant amplitude is the root of the sum of
the squares of the original amplitudes. Put more briefly – if the
frequencies are very different, the total intensity is simply given by the
sum of the two constituent intensities.
These fluctuations are known as ‘beats’, and the difference f2-f1 is known
as the beat frequency. To give an illustration: While tuning a violin, if the
tuning is slightly off-key, you will hear the note pulse: loud-soft-loud-soft
and so on. As you get closer to the correct note, the pulsing slows down
until, when the instrument is in tune, no pulsing is heard at all because
f2-f1=0.
4.2.8
Adding more than two waves
4.2.8.1 Diffraction Grating
The first case we come to with more than two waves is the diffraction
grating. This is a plate with many narrow transparent regions. The light
can only get through these regions. If the distance between adjacent
‘slits’ is d, we obtain constructive interference, as in section 4.2.4.1,
Page 62
Corrections March 2007
when d sin   n - in other words when the light from all slits is in
phase.
The difference between this arrangement and the double-slit is that
when d sin   n we find that interference is more or less destructive.
Therefore a given colour (or wavelength) only gets sent in particular
directions. We can use the device for splitting light into its constituent
colours.
4.2.8.2 Bragg Reflection
A variation on the theme of the diffraction grating allows us to measure
the size of the atom.

d

The diagram shows a section of a crystal. Light (in this case, X-rays) is
bouncing off the layers of atoms. There are certain special angles for
which all the reflections are in phase, and interfere constructively.
Looking at the small triangle in the diagram, we see that the extra path
travelled by the wave bouncing off the second layer of atoms is
D  2d sin  .
(20)
When D=n, we have constructive interference, and a strong reflection.
There is one thing that takes great care – notice the definition of q in the
diagram. It is not the angle of incidence, nor is it the angle by which the
ray is deflected – it is the angle between surface and ray. This is equal
to half the angle of deflection, also equal to /2 – i.
Using this method, the spacing of atomic layers can be calculated – and
this is the best measurement we have for the ‘size’ of the atom in a
crystal.
4.2.8.3 Diffraction
What happens when we add a lot of waves together? There is one case
we need to watch out for – when all possible phases are represented
with equal strength. In this case, for each wave cost    , there will be
an equally strong wave cost        cost    , which will cancel it
out.
Page 63
Corrections March 2007
How does this happen in practice? Look at the diagram below.
Compared with the wave from the top of the gap, the path differences of
the waves coming from the other parts of the gap go from zero to
Dmax  W sin  , where W is the width of the gap.
To obser vation
point

W

Dmax = W sin 
If kDmax is a multiple of 2, then we will have all possible phases
represented with equal strength, and overall destructive interference will
result.
To summarize, destructive interference is seen for angles , where
W sin   n .
(21)
Make sure you remember that W is the width of the gap, and that this
formula is for destructive interference.
This formula is only valid (as in the diagram) when the observer is so far
away that the two rays drawn are effectively parallel. Alternatively the
formula works perfectly when it is applied to an optical system that is
focused correctly, for then the image is at infinity.
4.2.8.4 Resolution of two objects
How far away do you have to get from your best friend before they look
like Cyclops? No offence – but how far away do you have to be before
you can’t tell that they’ve got two eyes rather than one? The results of
diffraction can help us work this out. Let’s call this critical distance L.
The rays from both eyes come into your eyeball. Let us suppose that
the angle between these rays is , where  is small, and that your
friend’s eyes are a distance s apart. Therefore tan   sin   s L .
These two rays enter your eye, and spread out (diffract) as a result of
passing through the gap called your pupil. They can only just be
‘resolved’ – that is noticed as separate – when the first minimum of one’s
diffraction pattern lines up with the maximum of the other. Therefore
W sin    where W is the width of your pupil.
Page 64
Corrections March 2007
Pupil of your eye

Putting the two formulae together gives:
sin  
L
s 

L W
.
sW
(22)

For normal light (average wavelength about 500nm), a 5mm pupil, and a
10cm distance between the eyes: you friend looks like Cyclops if you are
more than 1km away! If you used a telescope instead, and the
telescope had a diameter of 10cm, then your friend’s two eyes can be
distinguished at a distances up to 20km.
4.2.8.5 The Bandwidth Theorem
In the last section, we asked the question, “What happens when you add
lots of waves together?” However we cheated in that we only
considered waves of the same frequency. What happens if the waves
have different frequencies?
Suppose that we have a large number of waves, with frequencies evenly
spread between f and f+f. The angular frequencies will be spread from
 to +, where =2f as in equation (4). Furthermore, imagine that
we set them up so that they all agree in phase at time t=0. They will
never agree again, because they all have different frequencies.
The phases of the waves at some later time t will range from t to
()t.
Initially we have complete constructive interference. After a short time
t, however, we have destructive interference. This will happen when
(as stated in the last section) all phases are equally represented – when
the range of phases is a whole multiple of 2. This happens when
t×=2. After this, the signal will stay small, with occasional complete
destructive interference.
From this you can reason (if you’re imaginative or trusting) that if you
need to give a time signal, which has a duration smaller than t, you
must use a collection of frequencies at least =2/t. This is called the
bandwidth theorem. This can be stated a little differently:
Page 65
Corrections March 2007
 t  2
 2f t  2 .
f t  1
(23)
A similar relationship between wavelength and length can be obtained, if
we allow the wave to have speed c:
 x
c
f     1
f
c
1
   x  1

   x  1 .
(24)
Expressed in terms of the wave number k, this becomes:
 k 
 x  1
.
 2 
k x  2

(25)
In other words, if you want a wave to have a pulse of length x at most,
you must have a range of k values of at least 2/x.
4.2.8.6 Resolution of spectra
A spectrometer is a device that measures wavelengths. Equation (25)
can be used to work out the accuracy (or resolution) of the
measurement.
If you want a minimum error k in the wavenumber, you must have a
distance of at least x=2/k. But what does this distance mean? It
transpires that this is the maximum path difference between two rays in
going through the device – and as such is proportional to the size of the
spectrometer.
So, the bigger the spectrometer, the better its
measurements are.
4.2.9
Doppler Effect
4.2.9.1 Classical Doppler Effect
Suppose a bassoonist is playing a beautiful pure note with frequency f.
Now imagine that he is practising while driving along a road. A fellow
motorist hears the lugubrious sound. What frequency does the listener
hear? Let us suppose that the player is moving at velocity u, and the
listener is moving at velocity v. For simplicity we only consider the
problem in one dimension, however velocities can still be positive or
negative.
Page 66
Corrections March 2007
Furthermore, imagine that the distance between player and listener is L0
at time zero, when the first wave-peak is broadcast from the bassoon.
We assume that the waves travel at speed c with respect to the ground.
This peak is received at time t1, where
L0  vt1  ct1
L0  c  v t1
(26)
The first line is constructed like this: The travellers start a distance L0
apart, so by the time the signal is received, the distance between them is
L0 + vt1. This distance is covered by waves of speed c in time t1 – hence
the right hand side.
The next wave peak will be broadcast at time 1/f – one wave cycle later.
At this time, the distance between the two musicians will be
L0  v  u T  L0  v  u  / f . This second peak will be received at time
t2, where
L0 


vu
1
1
 v t 2    c t 2  
f
f
f


.
cu
L0 
 c  v t 2
f
(27)
Finally we can work out the time interval elapsed between our listener
hearing the two peaks, and from this the apparent frequency is easy to
determine.
t 2  t1 c  v   L0  c  u  L0
f
cu
1
 t 2  t1  
f
f c  v 
cv
f f 
cu
(28)
From this we see that if v=u, no change is observed. If the two are
approaching, the apparent frequency is high (blue-shifted). If the two are
receding, the apparent frequency is low (red-shifted).
4.2.9.2 Relativistic Doppler Effect
Please note that if either u or v are appreciable fractions of the speed of
light, this formula will give errors, and the relativistic calculation must be
used.
For light only, the relativistic formula is
Page 67
Corrections March 2007
cu
,
cu
f f
(29)
where u is the approach velocity as measured by the observer (u is
negative if the source and observer are receding). The relativistic form
for other waves is more complicated, and will be left for another day.
4.3 Rays
4.3.1
Reflection and Refraction
All the discussion so far has centred on the oscillatory nature of waves.
We can predict some of the things waves do without worrying about the
oscillations – like reflection and refraction. The diagram below shows
both. We refer to a refractive index of a material, which is defined as
Refractive Index ( n) 
Speed of light in vacuum
.
Speed of light in the material
(30)
Air has a refractive index of about 1.0003 14 , glass has a refractive index
of about 1.5, and water about 1.3.
i
i
n1
n2
r
First of all, the angle of reflection is equal to the angle of incidence (both
were labelled i in the diagram).
Secondly, the angle of incidence is related to the angle of refraction r by
the formula:
sin i n2

sin r n1
14
(31)
The refractive index also gives a measure of pressure, since n-1 is proportional to pressure.
Page 68
Corrections March 2007
Notice that since the sine of an angle can be no larger than one, if n2<n1,
then refraction becomes impossible if i  sin 1 n2 n1  . This limiting angle
is called the critical angle. For greater angles of incidence, the entire
wave is reflected, and this is called total internal reflection.
When a wave passes from one material into another, the frequency
remains the same (subject to the linearity provisos of section 4.1.1).
Given that the speed changes, the wavelength will change too. The
wavelength of light in a particular material can be evaluated:

c c 0 0


f nf
n
(32)
where c0 (and 0) represent speed of light (and wavelength) in vacuum.
4.4 Fermat’s Principle
Fermat’s principle gives us a method of working out the route light will
take in an optical system. It states that light will take the route that takes
the least time. Given that the time taken in a single material is equal to
T
D nD

c
c0
(33)
where c0 is the speed of light in vacuum; minimizing the time is the same
as minimizing the product of distance and refractive index. This latter
quantity (nD) is called the optical path. It is possible to prove the laws of
reflection and refraction using this principle. 15
4.5 Questions
1.
A hole is drilled through the Earth from the U.K. to the centre of the Earth
and out of the other side. All the material is sucked out of it, and a 1kg
mass is dropped in at the British end. How much time passes before it
momentarily comes to rest at the Australian end? (NB You may need
some hints from section 1.2.4) +
2.
Repeat q1 where a straight hole is drilled between any two places on
Earth. Assume that the contact of the mass with the sides of the hole is
frictionless. ++
15
To do this, imagine the plane as a sheet of graph paper, with the boundary along the x-axis.
Suppose that the light starts at point (0,Y), and needs to get to (X,-Y). Now assume that the
light crosses the x-axis at point (x,0). Work out the total optical path travelled along the route,
and then minimize it with respect to x. You should then be able to identify sin i and sin r in the
algebraic soup, and from this, you should be able to finish the proof.
Page 69
Corrections March 2007
3.
Show that Fermat’s principle allows you to ‘derive’ the Law of Reflection.
Assume that you have a mirror along the x-axis. Let light start at point
(X1,Y1) and end at point (X2,Y2). Show that the least-time reflected route
is the one which bounces off the mirror where angles of incidence and
reflection will be equal. +
4.
Show that Fermat’s principle allows you to ‘derive’ Snell’s Law. Assume
that you have a material with refractive index 1 for y>0 (that is, above the
x-axis), and refractive index n where y<0. Show that the shortest time
route from point (X1,Y1) to (X2,Y2), where Y1>0 and Y2<0, crosses the
boundary at the point where sin i / sin r = n. +
5.
You are the navigator for a hiking expedition in rough ground. Your
company is very thirsty and tired, and your supplies have run out. There
is a river running East-West which is 4km South of your current position.
Your objective is to reach the base camp (which is 2km South and 6km
West of your current position), stopping off at the river on the way. What
is the quickest route to the camp via the river? +
6.
You are the officer in charge of a food convoy attempting to reach a
remote village in a famine-stricken country. On your map, you see that
50km to your East is a straight border (running North-South) between
scrub land (over which you can travel at 15km/hr) and marsh (over which
you can only travel at 5km/hr). The village is 141km South-East of your
current position. What is the fastest route to reach the village? +
7.
In an interferometer, a beam of coherent monochromatic light (with
wavelength ) is split into two parts. Both parts travel for a distance L
parallel to each other. One travels in vacuum, the other in air. The
beam is then re-combined. If destructive interference results, what can
you say about L, and nair?
8.
Your wind band is about to play on a pick-up truck going down a
motorway at 30m/s. You want people on the bridges overhead to hear
you playing ‘in tune’ (such that treble A is 440Hz) when you are coming
directly towards them.
What frequency should you tune your
instruments to?
9.
A police ‘speed gun’ uses microwaves with a wavelength of about 3cm.
The ‘gun’ consists of a transmitter and receiver, with a small mirror which
sends part of the transmission directly into the receiver. Here it
interferes with the main beam which has reflected off a vehicle. The
received signal strength pulsates (or beats). What will be the frequency
of this pulsation if the vehicle is travelling towards you at 30mph? +
10. How far away do you need to hold a ruler from your left eye before you
can no longer resolve the millimetre markings? Keep your right eye
covered up during this experiment. Use equation (22) to make an
Page 70
Corrections March 2007
estimate for the wavelength of light based on your measurement.
Remember that W is the width of your pupil.
11. A signal from a distant galaxy has one third of the frequency you would
expect from a stationary galaxy. Calculate the galaxy’s recession
velocity using equation (29), and comment on your answer. (NB redshifts this big are measured with very distant astronomical objects.)
Page 71
Corrections March 2007
5 Hot Physics
This section gives an introduction to the areas of physics known as
thermodynamics and statistical mechanics.
These deal with the
questions “What happens when things heat up or cool down?” and
“Why?” respectively.
We start with a statement that will be very familiar – but then find that it
leads us into new territory when explored further.
5.1 The Conservation of Energy
You will be used to the idea that energy can neither be used up nor
created – only transferred from one object to another, perhaps in
different forms.
For our purposes, this is stated mathematically as
dQ + dW = dU ,
(1)
where ‘dX’ refers to ‘a small change in X’. Put into words, this states:
“Heat entering object + Work done on the object = the change in its
internal energy.” Internal energy means any form of stored energy in the
object. Usually this will mean the heat it has, and will be measured by
temperature. However if magnetic or electric fields are involved, U can
also refer to electrical or magnetic potential energy.
Given that the conservation of energy must be the starting point for a
study of heat, it is called the First Law of Thermodynamics.
Equation (1) can be applied to any object or substance. The most
straightforward material to think about is a perfect gas, and so we shall
start there. It is possible to generalize our observations to other
materials afterwards.
Imagine some gas in a cylinder with a piston of cross-sectional area A.
The gas will have a volume V, and a pressure p. Let us now do some
work on the gas by pushing the piston in by a small distance dL. The
force required to push the piston F = p A, and so the work done on the
gas is dW = F dL = p A dL . Notice that AdL is also the amount by which
the volume of the gas has been decreased. If dV represents the change
in volume, dV = -A dL. Therefore dW = -p dV.
For a perfect gas in a cylinder (or in fact in any other situation), the First
Law can be written a bit differently as:
dQ = dU + p dV
Page 72
(2)
Corrections March 2007
5.2 The Second Law
While the First Law is useful, there are certain things it can never tell us.
For example – think about an ice cube sitting on a dish in an oven. We
know what happens next – the ice cube melts as heat flows from oven to
ice, warming it up until it reaches melting point. However the First Law
doesn’t tell us that. As far as it is concerned it is just as possible for heat
to flow from the ice to the oven, cooling the ice and heating the oven.
We stumbled across our next law – called the second law of
thermodynamics. This can be stated in several ways, but we shall start
with this: Heat will never flow from a cold object to a hotter object by
itself.
This helps us with the ice in the oven, but you may be wondering what
the significance of the “by itself” is. Actually heat can be transferred from
a cold object to a hotter one – that is what fridges and air conditioning
units do. However they can only do it because they are plugged into the
electricity supply. If you are prepared to do some work – then you can
get heat out of a cold object and into a hotter one, but as soon as you
turn the power off and leave it to its own devices, the heat will start
flowing the other way again.
5.3 Heat Engines and Fridges
Hot ‘reservoir’ - the
atmosphere at temperature
T(h)
dQ1
Work dW provided by
electric compressor
Fridge
dQ2
Cold ‘reservoir’ - the ice
box at temperature T(c)
The fridge is shown diagrammatically above. It is a device which uses
work dW (usually provided by an electric compressor) to extract heat
dQ2 from the ice-box (cooling it down), and pump it out into the
surroundings (warming them up). However, by the conservation of
energy, the amount of energy pumped out dQ1 is bigger than the amount
of energy removed from the ice-box. By convention dQ2>0, and dQ1<0,
since heat flowing in is regarded as positive. The First Law therefore
states that dQ1 + dQ2 + dW = 0.
Page 73
Corrections March 2007
The fridge is a device that uses work to move heat from cold objects to
hot. The opposite of a fridge is a heat engine. This allows heat to flow
its preferred way – namely from hot to cold – but arranges it to do some
work on the way. Petrol engines, steam engines, turbo-generators and
jet engines are all examples of heat engines.
Hot ‘reservoir’ - the fire-box
at temperature T(h)
dQ1
Work dW produced,
drives electrical
generator
Engine
dQ2
Cold ‘reservoir’ - the
atmosphere at temperature
T(c)
It was Carnot who realised that the most efficient heat engine of all was
a ‘reversible’ heat engine. In other words – one that got the same
amount of work out of the heat transfer as would be needed to operate a
perfect fridge to undo its operation.
In order to do this, it is necessary for all the heat transfers (between one
object and another) to take place with as small a temperature difference
as possible. If this is not done, heat will flow from hot objects to cold – a
process which could have been used to do work, but wasn’t. Therefore
not enough work will be done to enable the fridge to return the heat to
the hot object.
Carnot therefore proposed that the ratio of heat coming in from the hot
object to the heat going out into the cold object has a maximum for this
most-efficient engine. This is because the difference between heat in
and heat out is the work done, and we want to do as much work as
possible. Furthermore, he said that this ratio must be a function of the
temperatures of the hot and cold objects only.
This can be stated as
dQ1
 f (T1 , T2 )
dQ2
(3)
where T1 is the temperature of the hot object, and T2 is that of the cold
object. More light can be shed on the problem if we stack two heat
engines in series, with the second taking the heat dQ2 from the first (at
Page 74
Corrections March 2007
temperature T2), extracting further work from it before dumping it as heat
(dQ3) into a yet colder object at temperature T3.
The two heat engines separately and together give us the equations:
dQ1
 f (T1 , T2 )
dQ2

dQ2
 f (T2 , T3 )
dQ3
f (T1 , T3 )  f (T1 , T2 )  f (T2 , T3 )

5.3.1
dQ1
 f (T1 , T3 )
dQ3
f (T1 , T2 ) 
.
(4)
g (T1 )
g (T2 )
Thermodynamic Temperature
However if g(T) is a function of the temperature alone, we might as well
call g(T) the temperature itself. This is the thermodynamic definition of
temperature.
To summarize: Thermodynamic temperature (T) is defined so that in a
reversible heat engine, the ratio of heat extracted from the hot object
(Q1) to the heat ejected into the cold object (Q2):
Q1 T1

Q 2 T2
(5)
The ‘kelvin’ temperature scale obtained using the gas laws satisfies this
definition. For this reason, the kelvin is frequently referred to as the unit
of ‘thermodynamic temperature’.
5.3.2
Efficiency of a Heat Engine
The efficiency of a reversible heat engine can then be calculated. We
define the efficiency () to be the ratio of the work done (the useful
output) to Q1 (the total energy input). Therefore

dQ1  dQ2
T
dW

1 2 .
dQ1
dQ1
T1
(6)
This, being the efficiency of a reversible engine, is the maximum
efficiency that can be achieved. A real engine will fall short of this goal.
Notice that for a coal-fired power station, in which T1 (the temperature of
the boiler) is frequently 840K, and T2 (the temperature of the stream
outside) is 300K; the maximum possible efficiency is
 1
300
 64% .
840
Page 75
Corrections March 2007
In practice the water leaves the turbo generator at 530K, and so the
efficiency can’t go any higher than
530
 37% .
840
 1
The design of modern large power stations is such that the actual
efficiency is remarkably close to this value.
5.4 Entropy
Now we need to take a step backwards before we can go forwards.
Look back at the definition of thermodynamic temperature in equation
(5). It can be rearranged to state
dQ1 dQ2

T1
T2

dQ1 dQ2

0
T1
T2
REVERSIBLE.
(7)
Remember that this is for the ideal situation of a reversible process – as
in a perfect fridge or heat engine. Suppose, then, that we start with
some gas at pressure p and volume V. Then we do something with it
(squeeze it, heat it, let it expand, or anything reversible), and finally do
some more things to it to bring it back to pressure p and volume V. The
list of processes can be broken up into tiny stages, each of which saw
some heat (dQ) entering or leaving the system, which was at a particular
temperature T. The only difference between this situation, and that in (7)
is that there were only two stages in the process for the simpler case.
The physics of (7) should still apply, no matter how many processes are
involved. Therefore providing all the actions are reversible we can write
dQ
0 
complete cycle T


dQ
 0 REVERSIBLE
T
(8)
where the circle on the integral implies that the final position (on a p,V
graph) is the same as where the gas started.
Now suppose that there are two points on the (p,V) graph which are of
interest to us, and we call them A and B. Let us go from A to B and then
back again (using a different route), but only using reversible processes.
We call the first route I, and the second route II. Equation (8) tells us
Page 76
Corrections March 2007
dQ B dQ I A dQ II
 T  A T  B T  0
dQ I B dQ II
A T  A T  0
B
REVERSIBLE
(9)
B
dQ I B dQ II
A T  A T
In other words the integral of dQ/T between the two points A and B is the
same, no matter which reversible route is chosen. This is a very special
property of a function – we label dQ/T as a function of state, and call it
the entropy (S).
This means that the current entropy of the gas, like pressure, volume
and temperature, is only a function of the state that the gas is in now –
and does not depend on the preparation method.
5.5 Irreversible Processes and the Second Law
We must stress that entropy is only given by  dQ T when the integral is
taken along reversible processes in which there is no wastage of heat.
Heat is wasted when it is allowed to flow from a hot object to a cold one
without doing any work on the journey. This would be irreversible, since
you could only get the heat back into the hot object if you expended
more energy on it.
Let us make an analogy. Reversible processes are like a world in which
purchasing prices and selling prices are the same. If you started with
£100, and spent it in various ways, you could sell the goods and end up
with £100 cash at the end.
Irreversible processes are like the real world in that a trader will want to
sell you an apple for more than she bought it for. Otherwise she won’t
be able to make a profit. If you started with £100, and spent it, you
would never be able to get the £100 back again, since you would lose
money in each transaction. You may end up with £100 ‘worth of goods’,
but you would have to be satisfied with a price lower than £100 if you
wanted to sell it all for cash.
Let us now return to the physics, and the gas in the piston. What does
irreversibility mean here? We haven’t lost any energy – the First Law
has ensured that. But we have lost usefulness.
Equation (8) tells us that if we come back to where we started, and only
use reversible processes on the way, the total entropy change will be
zero. There is another way of looking at this, from the point of view of a
heat engine.
Page 77
Corrections March 2007
Let us suppose that the temperature of the boiler in a steam engine is
TA. In a perfect heat engine, the cylinder will receive the steam at this
temperature. Suppose Q joules of heat are transferred from boiler to
cylinder. The boiler loses entropy Q/TA, the cylinder gains entropy Q/TA,
and the total entropy remains constant.
Now let us look at a real engine. The boiler must be hotter than the
cylinder, or heat would not flow from boiler to cylinder! Suppose that the
boiler is still at TA, but the cylinder is at TC. We have now let
irreversibility loose in the system, since the heat Q now flows from hot to
cooler without doing work on the way.
What about the entropy? The boiler now loses Q/TA to the connecting
pipe 16 , but the cylinder gains Q/TC from it. Since TC<TA, the cylinder
gains more entropy than the boiler lost.
This is an alternative definition of the Second Law. Processes go in the
direction to maximize the total amount of entropy.
5.6 Re-statement of First Law
For reversible processes, dW = -p dV, and dQ = T dS. Therefore the
First Law (1), can be written as
T dS = dU + p dV .
(9)
We find that this equation is also true for irreversible processes. This is
because T, S, U, p and V are all functions of state, and therefore if the
equation is true for reversible processes, it is true for all processes.
However care must be taken when using it for irreversible processes,
since TdS is no longer equal to the heat flow, and pdV is no longer equal
to the work done.
5.7 The Boltzmann Law
The Boltzmann Law is simple to state, but profound in its implications.
Probability that a particle has energy E  e  E kT
16
(10)
What’s the pipe got to do with it? Remember that we said that change in entropy dS is only
given by dQ/T for reversible processes. The passing of Q joules of heat into the pipe is done
reversibly (at temperature TA), so we can calculate the entropy change. Similarly the passing
of Q joules of heat from pipe to cylinder is done reversibly (at TC), so the calculation is
similarly valid at the other end. However something is going on in the pipe which is not
reversible – namely Q joules of heat passing from higher to lower temperature. Therefore we
mustn’t apply any dS = dQ/T arguments inside the pipe.
Page 78
Corrections March 2007
where k is the Boltzmann constant, and is about 1.38×10-23 J/K. We
also find that the probability that a system has energy E or greater is
also proportional to e-E/kT (with a different constant of proportionality).
There is common sense here, because (10) is saying that greater
energies are less likely; and also that the higher the temperature, the
more likely you are to have higher energies.
Let’s give some examples:
5.7.1
Atmospheric Pressure
The pressure in the atmosphere at height h is proportional to the
probability that a molecule will be at that height, and is therefore
proportional to e-mgh/kT. Here, the energy E, is of course the gravitational
potential energy of the molecule – which has mass m.
The proof of this statement is in several parts. Firstly we assume that all
the air is at the same temperature. This is a dodgy assumption, but we
shall make do with it. Next we divide the atmosphere into slabs (each of
height dh and unit area), stacked one on top of the other. Each slab has
to support all the ones above it. From the Gas Law (pV = NkT where N
is the number of molecules under consideration), and the definition of
density (Nm=V) we can show that =pm/kT. Furthermore, if you go up
by a small height dh, the pressure will reduce by the weight of one slab –
namely gdh. Therefore
dp
pmg
  g  
dh
kT
 mgh 
 p  p 0 exp 

 kT 
(11)
where p0 is the pressure at ground level. We see that the Boltzmann
Law is obeyed.
5.7.2
Velocity distribution of molecules in a gas
The probability that a molecule in the air will have x-component of its
velocity equal to ux is proportional to exp(-mux2/2kT). Here the energy E
is the kinetic energy associated with the x-component of motion.
From this statement, you can work out the mean value of ux2, and find it
to be:
 x exp mx 2kT dx  2kT m

 exp mx 2kT dx 2kT m
2
u
2
x
2
1
2
3/ 2
1/ 2
2
1
2
mu x2  12 kT
Page 79

kT
m
(12)
Corrections March 2007
The mean kinetic energy is given by


K  12 mu 2  12 m u x2  u y2  u z2  32 kT
(13)
so the internal energy of a mole of gas (due to linear motion) is
U  N A K  32 N A kT  32 RT
(14)
where R  N A k is the gas constant. From this it follows that the molar
heat capacity of a perfect gas 17 , CV  32 R .
5.7.3
Vapour Pressure
The probability that a water molecule in a mug of tea has enough (or
more than enough) energy to leave the liquid is proportional to exp(EL/kT) where EL is the energy required to escape the attractive pull of
the other molecules (latent heat of vaporization per molecule).
5.7.4
Justification of Boltzmann Law
In this section, we introduce some statistical mechanics to give a taste of
where the Boltzmann law comes from.
Suppose that you have N atoms, and P ‘packets’ of energy to distribute
between them. How will they be shared? In statistical mechanics we
assume that the energy will be shared in the most likely way.
In a simple example, we could try sharing 4 units of energy (P=4) among
Because the individual energy units are
7 atoms (N=7).
indistinguishable (as are the 7 atoms), the possible arrangements are:





1 atom with 4 energy units, 6 atoms with none
1 atom with 3 energy units, 1 with 1, 5 with none
2 atoms with 2 energy units, 5 with none
1 atom with 2 energy units, 2 with 1, 4 with none
4 atoms with 1 energy unit, 3 with none.
These are said to be the five macrostates of the system. We can work
out how likely each one is to occur if the energy is distributed randomly
by counting the ways in which each macrostate could have happened.
17
This is the heat capacity due to linear motion. For a monatomic gas (like helium), this is the
whole story. For other gases, the molecules can rotate or vibrate about their bonds as well,
and therefore the heat capacity will be higher.
Page 80
Corrections March 2007
For example, in the first case (all of the energy is given to one of the
atoms), there are seven ways of setting it up – because there are seven
atoms to choose from.
In the second case, we have to choose one atom to take 3 units (7 to
choose from), and then choose one from the remaining six to take the
remaining unit. Therefore there are 76 = 42 ways of setting it up.
Similarly we can count the ways of rearranging for the other
macrostates 18 :





1 atom with 4 energy units, 6 atoms with none
1 atom with 3 energy units, 1 with 1, 5 with none
2 atoms with 2 energy units, 5 with none
1 atom with 2 energy units, 2 with 1, 4 with none
4 atoms with 1 energy unit, 3 with none.
7 ways
42 ways
21 ways
105 ways
35 ways
We can see that there is one macrostate clearly in the lead – where 4
atoms have no energy, 2 atoms have one energy unit, and 1 atom has
two energy units. This macrostate is interesting because there is a
geometric progression in the number of atoms (4,2,1) having each
amount of energy.
It can be shown that the most likely macrostate will always be the one
with (or closest to) a geometric progression of populations. 19 In other
18
The calculation is made more straightforward using the formula W  N ! n0 ! n1! n2 !
where W is the number of ways of setting up the macrostate, n0 is the number of atoms with
no energy, n1 is the number of atoms with one unit of energy, and so on.
Here is one justification for this formula: N! gives the total number of ways of choosing the
atoms in order. The division by n0! prevents us overcounting when we choose the same
atoms to have zero energy in a different order. A similar reason holds for the other terms on
the denominator.
Alternatively, W = the number of ways of choosing n0 atoms from N  the number of ways of
choosing the n1 from the remaining (Nn0)  the number of ways of choosing the n2 from the
remaining (Nn0n1) and so on. Thus
W  C nN0 C nN1n 0 C nN2n 0n1 

19
N  n0 !  N  n0  n1 !  
N!
N!

n0 ! N  n0 ! n1!  N  n0  n1 ! n2 !  N  n0  n1  n2 !
n0 ! n1! n2 !
To show this, start with a geometric progression (in other words, say ni  Af
i
where f is
some number), and write out the formula for W. Now suppose that one of the atoms with i
units of energy gives a unit to one of the atoms with j units. This means that ni and nj have
gone down by 1, while ni1 and nj+1 have each gone up by one. By comparing the old and
Page 81
Corrections March 2007
words, assuming that this macrostate is the true one (which is the best
bet) 20 , then the fraction of atoms with n energy units is proportional to
some number to the power n. The actual fraction is given by the
1
n
formula 1  P N   1  N P  . Assuming that the mean number of
energy units per atom is large (so that we approximate the continuous
range of values that the physical energy can take), this means that
1  N P  e N P , and so the probability that an atom will have n units of
energy is proportional to e  Nn P .
Suppose that each packet contains  joules of energy. Then the energy
of one atom (with n packets) is E = n, and the probability that our atom
will have energy E as e  NE P . The mean energy per atom is P. Now
we have seen that the average energy of an atom in a system is about
kT, where T is the temperature in kelvins, and so it should not seem odd
that the Boltzmann probability is e  E kT where we replace one expression
for the mean energy per atom P, with another kT.
5.8 Perfect Gases
All substances have an equation of state. This tells you the relationship
between volume, pressure and temperature for the substance. Most
equations of state are nasty, however the one for an ideal, or perfect,
gas is straightforward to use. It is called the Gas Law. This states that
pV=nRT
(18)
pV=NkT
(19)
where p is the pressure of the gas, V its volume, and T its absolute (or
thermodynamic) temperature. This temperature is measured in kelvins
always. There are two ways of stating the equation: as in (18), where n
represents the number of moles of gas; or as in (19), where N
represents the number of molecules of gas. Clearly N=NAn where NA is
the Avogadro number, and therefore R=NAk.
You can adjust the equation to give you a value for the number density
of molecules. This means the number of molecules per cubic metre, and
is given by N/V = p/kT. The volume of one mole of molecules can also
be worked out by setting n=1 in (18):
new values of W, you can show that the new W is smaller, and that therefore the old
arrangement was the one with the biggest W.
20
The bet gets better as the number of atoms increases. The combination (4,2,1) was the
most popular in our example of N=7, P=4, however if you repeated the exercise with N=700
and P=400, you would find a result near (400,200,100) almost a certainty. In physics we deal
with huge numbers of atoms in matter, so the gambling pays off.
Page 82
Corrections March 2007
Vm 
RT
p
.
(20)
You can adjust this equation to give you an expression for the density. If
the mass of one molecule is m, and the mass of a mole of molecules
(the R.M.M.) is M, we have

Mp
Mass
M


Volume RT p RT
N mp mp
 A

N A kT kT
.
(21)
Please note that this is the ideal gas law. Real gases will not always
follow it. This is especially true at high pressures and low temperatures
where the molecules themselves take up a good fraction of the space.
However at room temperature and atmospheric pressure, the Gas Law
is a very good model.
5.8.1
Heat Capacity of a Perfect Gas
We have already shown (in section 5.7.2) that for a perfect gas, the
internal energy due to linear motion is 32 RT per mole. If this were the
only consideration, then the molar heat capacity would be 32 R . However
there are two complications
5.8.1.1 The conditions of heating
In thermodynamics, you will see molar heat capacities written with
subscripts – CP and CV. They both refer to the energy required to heat a
mole of the substance (M kilograms) by one kelvin. However the energy
needed is different depending on whether the volume or the pressure is
kept constant as the heating progresses.
When you heat a gas at constant volume, all the heat going in goes into
the internal energy of the gas (dQV = dU).
When you heat a gas at constant pressure, two things happen. The
temperature goes up, but it also expands. In expanding, it does work on
its surroundings. Therefore the heat put in is increasing both the internal
energy and is also doing work (dQP = dU + p dV).
Given that we know the equation of state for the gas (18), we can work
out the relationship between the constant-pressure and constant-volume
heat capacities. In these equations we shall be considering one mole of
gas.
Page 83
Corrections March 2007
dQV dU

dT
dT
dQ P dU
dV
CP 

p
dT
dT
dT
d  RT 

  CV  R
 CV  p
dT  p 
CV 
(22)
5.8.1.2 The type of molecule
Gas molecules come in many shapes and sizes. Some only have one
atom (like helium and argon), and these are called monatomic gases.
Some gases are diatomic (like hydrogen, nitrogen, oxygen, and
chlorine), and some have more than two atoms per molecule (like
methane).
The monatomic molecule only has one use for energy – going places
fast. Therefore its internal energy is given simply by 32 kT , and so the
molar internal energy is U  32 RT . Therefore, using equation (22), we
can show that CV  32 R and C P  CV  R  52 R .
A diatomic molecule has other options open to it. The atoms can rotate
about the molecular centre (and have a choice of two axes of rotation).
They can also wiggle back and forth – stretching the molecular bond like
a rubber band. At room temperature we find that the vibration does not
have enough energy to kick in, so only the rotation and translation (the
linear motion) affect the internal energy.
Each possible axis of rotation adds 12 kT to the molecular energy, and so
we find that for most diatomic molecules, CV  52 R and C P  72 R .
5.8.1.3 Thermodynamic Gamma
It turns out that the ratio of C P CV crops up frequently in equations, and
is given the letter . This is not to be confused with the  in relativity,
which is completely different.
Using the results of our last section, we see that =5/3 for a monatomic
gas, and =7/5 for one that is diatomic.
5.8.2
Pumping Heat
If a healthy examiner expects you to know about ideal gases and
thermodynamics – you can bet that he or she will want you to be able to
do thermodynamics with an ideal gas. In this section we show you how
to turn a perfect gas (in a cylinder) into a reversible heat engine, and in
doing so we will introduce the techniques you need to know.
Page 84
Corrections March 2007
5.8.2.1 Isothermal Gas Processes
As an introduction, we need to know how to perform two processes.
Firstly we need to be able to get heat energy into or out of a gas without
changing its temperature. Remember that we want a reversible heat
engine, and therefore the gas must be at the same temperature as the
hot object when the heat is passing into it. Any process, like this, which
takes place at a constant temperature is said to be isothermal.
The Gas Law tells us (18) that pV=nRT, and hence that pV is a function
of temperature alone (for a fixed amount of gas). Hence in an isothermal
process
pV  const .
ISOTHERMAL
(23)
Using this equation, we can work out how much we need to compress
the gas to remove a certain quantity of heat from it. Alternatively, we
can work out how much we need to let the gas expand in order for it to
‘absorb’ a certain quantity of heat. These processes are known as
isothermal compression, and isothermal expansion, respectively.
Suppose that the volume is changed from V1 to V2, the temperature
remaining T. Let us work out the amount of heat absorbed by the gas.
First of all, remember that as the temperature is constant, the internal
energy will be constant, and therefore the First Law may be stated
dQ=pdV. In other words, the total heat entering the gas may be
calculated by integrating pdV from V1 to V2:
Q   pdV  
V
nRT
V2
dV  nRT ln V V 1  nRT ln 2 .
V1
V
(24)
This equation describes an isothermal (constant temperature) process
only. In order to keep the temperature constant, we maintain a good
thermal contact between the cylinder of gas and the hot object (the boiler
wall, for example) while the expansion is going on.
5.8.2.2 Adiabatic Gas Processes
The other type of process you need to know about is the adiabatic
process. These are processes in which there is no heat flow (dQ=0),
and they are used in our heat engine to change the temperature of the
gas in between its contact with the hot object and the cold object.
Sometimes this is referred to as an isentropic process, since if dQ=0 for
a reversible process, TdS=0, and so dS=0 and the entropy remains
unchanged. 21
21
While the terms ‘isentropic’ and ‘adiabatic’ are synonymous for a perfect gas, care must be
taken when dealing with irreversible processes in more advanced systems. In this context dQ
is not equal to TdS. If dQ=0, the process is said to be adiabatic: if dS=0, the process is
Page 85
Corrections March 2007
Before we can work out how much expansion causes a certain
temperature change, we need to find a formula which describes how
pressure and volume are related in an adiabatic process. Firstly, the
First Law tells us that if dQ=0, then 0 = dU + p dV. We can therefore
reason like this for n moles of gas:
0  dU  pdV
 nCV dT  pdV
Now for a perfect gas, nRT=pV, therefore nRdT  pdV  Vdp . So we
may continue the derivation thus:
CV
 pdV  Vdp 
R
C
0  V  pdV  Vdp   pdV
R
 CV ( pdV  Vdp)  RpdV .
 C p pdV  CV Vdp
nCV dT 
(25)
  pdV  Vdp
dV dp


V
p
Integrating this differential equation gives
 ln V  ln p  C  0
pV   e C
. ADIABATIC
(26)
pV   const
Equation (26) is our most important equation for adiabatic gas
processes, in that it tells us how pressure and volume will be related
during a change.
We now come back to our original question: what volume change is
needed to obtain a certain temperature change? Let us suppose we
have a fixed amount of gas (n moles), whose volume changes from V1 to
V2. At the same time, the temperature changes from T1 to T2. We may
combine equation (26) with the Gas Law to obtain:
isentropic. Clearly for a complex system, the two conditions will be different. This arises
because in these systems, the internal energy is not just a function of temperature, but also of
volume or pressure.
Page 86
Corrections March 2007
pV   const
pV V  1  const
nRTV  1  const .
(27)
T1 V1 1
1
T2 V2 1
5.8.2.3 A Gas Heat Engine
We may now put our isothermal and adiabatic processes together to
make a heat engine. The engine operates on a cycle:
1.
The cylinder is attached to the hot object (temperature Thot), and
isothermal expansion is allowed (from V1 to V2) so that heat Q1 is
absorbed into the gas.
2.
The cylinder is detached from the hot object, and an adiabatic
expansion (from V2 to V3) is allowed to lower the temperature to
that of the cold object (Tcold).
3.
The cylinder is then attached to the cold object. Heat Q2 is then
expelled from the cylinder by an isothermal compression from V3 to
V4.
4.
Finally, the cylinder is detached from the cold object. An adiabatic
compression brings the volume back to V1, and the temperature
back to Thot.
Applying equation (24) to the isothermal processes gives us
Q1  nRThot ln
Q2  nRTcold
V2
V1
V
ln 4
V3
.
(28)
Similarly, applying equation (27) to the adiabatic processes gives us
V3  Thot

V2  Tcold



 1
 1
V4  Thot 


V1  Tcold 
V3 V4
V
V


 3  2
V2 V1
V4 V1
Combining equations (28) and (29), gives us
Page 87
.
(29)
Corrections March 2007
T
Q1
  hot
Q2
Tcold
T
Q1
 hot
Q2 Tcold
,
(30)
where the minus sign reminds us that Q2<0, since this heat was leaving
the gas.
To summarize this process, we have used a perfect gas to move heat
from a hot object to a colder one. In doing this, we notice less heat was
deposited in the cold object than absorbed from the hot one. Where has
it gone? It materialized as useful work when the cylinder was allowed to
expand. Had the piston been connected to a flywheel and generator, we
would have seen this in a more concrete way.
We also notice that we have proved that the kelvin scale of temperature,
as defined by the Gas Law, is a true thermodynamic temperature since
equation (30) is identical to (5).
5.9 Radiation of Heat
And finally... there is an extra formula that you will need to be aware of.
The amount of heat radiated from an object is given by:
P  AT 4
(31)
P is the power radiated (in watts), A is the surface area of the object (in
m2), and T is the thermodynamic temperature (in K).
The constant  is called the Stefan-Boltzmann constant, and takes the
value of 5.671×10-8 W/(m2K4).
The amount radiated will also depend on the type of surface. For a
perfect matt black (best absorber and radiator), the object would be
called a black body, and the emissivity  would take the value 1. For a
perfect reflector, there is no absorption, and no radiation either, so =0.
5.10
Questions
1.
Calculate the maximum efficiency possible in a coal fired power station,
if the steam is heated to 700°C and the river outside is at 7°C.
2.
Mechanical engineers have been keen to build jet engines which run at
higher temperatures. This makes it very difficult and expensive to make
the parts, given that the materials must be strong, even when they are
almost at their melting point. Why are they making life hard for
themselves?
Page 88
Corrections March 2007
3.
Two insulated blocks of steel are identical except that one is at 0°C,
while the other is at 100°C. They are brought into thermal contact. A
long time later, they are both at the same temperature. Calculate the
final temperature; the energy change and entropy change of each block
if (a) heat flows by conduction from one block to the other, and if (b) heat
flows from one to the other via a reversible heat engine. ++
4.
There is a ‘rule of thumb’ in chemistry that when you raise the
temperature by 10°C, the rate of reaction roughly doubles. Use
Boltzmann’s Law to show that this means the activation energy of
chemical processes must be of order 1019J. +
5.
The amount of energy taken to turn 1kg of liquid water at 100°C into 1kg
of steam at the same temperature is 2.26 MJ. This is called the latent
heat of vaporization of water. How much energy does each molecule
need to ‘free itself’ from the liquid?
6.
By definition, the boiling point of a liquid is the temperature at which the
saturated vapour pressure is equal to atmospheric pressure (about
100kPa). Up a mountain, you find that you can’t make good tea,
because the water is boiling at 85°C. What is the pressure? You will
need your answer to q4. +
7.
Estimate the altitude of the mountaineer in q5. Assume that all of the air
in the atmosphere is at 0°C. +
8.
Use the Gas Law to work out the volume of one mole of gas at room
temperature and pressure (25°C, 100kPa).
9.
What fraction of the volume of the air in a room is taken up with the
molecules themselves? Make an estimate, assuming that the molecules
are about 1010m in radius.
10. Estimate a typical speed for a nitrogen molecule in nitrogen at room
temperature and pressure. On average, how far do you expect it to
travel before it hits another molecule? Again, assume that the radius of
the molecule is about 1010m. ++
11. The fraction of molecules (mass m each) in a gas at temperature T
2
which have a particular velocity (of speed u) is proportional to e  mu / 2 kT ,
as predicted by the Boltzmann law. However the fraction of molecules
2
which have speed u is proportional to u 2 e  mu / 2 kT . Where does the u2
come from? ++
12. One litre of gas is suddenly squeezed to one hundredth of its volume.
Assuming that the squeezing was done adiabatically, calculate the work
done on the gas, and the temperature rise of the gas. Why is the
adiabatic assumption a good one for rapid processes such as this?
Page 89
Corrections March 2007
13. A water rocket is made using a 2 litre plastic drinks bottle. An amount of
water is put into the bottle, and the stopper is put on. Air is pumped into
the bottle through a hole in the stopper. When the pressure gets to a
certain level, the stopper blows out, and the pressure of the air in the
bottle expels the water. If the bottle was standing stopper-end
downwards, it flies up into the air. If you neglect the mass of the bottle
itself, what is the optimum amount of water to put in the bottle if you want
your rocket to (a) deliver the maximum impulse, or (b) rise to the
greatest height when fired vertically. ++
Page 90
Corrections March 2007
6 Sparks & Generation
6.1 Electrostatics – when things are still
The fundamental fact of electrostatics will be familiar to you – opposite
charges attract: like charges repel. As a physicist it is not enough to
know this, we also want to know how big the force is. It turns out that
the equations describing the force, energy, potential and so on are very
similar mathematically to the equations describing gravitational
attraction.
Using the symbols F for force, U for potential energy, V for potential and
E for field, we have the equations:
Qq
4 0 R 2
1 Q
Er 
4 0 R 2
1 Qq
U
4 0 R
Fr 
V
1
(1)
Q
4 0 R
1
Notice that the symbols are slightly different for gravity – field is now
given E rather than g, and in consequence we have to use another letter
for energy – hence our choice of U. Here we have put a charge Q at the
origin, and we measure the quantities associated with a small ‘test’
charge q at distance R. Notice that we do not have a minus sign in front
– this allows like charges to repel rather than attract (whereas in gravity,
positive mass attracts positive mass [and we have never found any
lumps of negative mass – if we did this would upset a lot of our
thinking]). Also, in place of the G of gravity, we have the constant
1 4 0 which is about 1020 times bigger. No wonder the theoreticians
talk of gravity as a weaker force!
Now, you may wonder, why the factor of 4? This comes about,
because the equations above are not the nicest way of describing
electrostatic forces. They are based on the ‘Coulomb Force Law’, which
is the first equation in (1) – however there is another, equivalent, way of
describing the same physics, and it is called Gauss’ Law of
Electrostatics. This Gauss Law is on the Olympiad syllabus, and you will
find it useful because it will simplify your electrical calculations a lot.
Page 91
Corrections March 2007
6.1.1
Gauss Law of Electrostatics
Firstly, let us say what the law is. Then we will describe it in words, and
then prove that it is equivalent to the Coulomb Law. Finally, we will
show its usefulness in other calculations.
Q
 E  dS  
S
(2)
0
What does this mean? Firstly let us look at the individual symbols. S is
a closed surface (that is what the circle on the integral sign means) – like
the outer surface of an apple, a table, or a doughnut – but not the outer
surface of a bowl (which ends at a rim). dS is a small part of the
surface, with area dS, and is a vector pointing outward, perpendicular to
the surface at that point. Q is the total electric charge contained inside
the surface S. Finally, the vector E is the electric field (in volts per metre)
– where the vector points in the direction a positive charge would be
pulled.
The odd looking integral tells us to integrate the dot product of E with dS
(a normal vector to the surface) around the complete surface. This may
sound very foreign, strange, and difficult, but let us give some examples.
First of all, suppose S is a spherical surface of radius R, with one charge
(+Q) at the centre of the surface. Assuming there are no other charges
nearby, the field lines will be straight, and will stream out radially from
the centre. Therefore E will be parallel to dS, and the dot product will
simply be E dS – the product of the magnitudes. Now the size of E must
be the same all round the surface, by symmetry. Therefore
 E  dS   E dS  E  dS  E 4R
S
S
2
(3)
S
since the surface area of a sphere is given by 4R2. Now by Gauss’
Law, this must equal Q  0 . Putting the two equations together gives us
4R 2 E 
Q
0

E
Q
4 0 R 2
(4)
in agreement with Coulomb’s Law.
Similarly we may find the field at a distance R from a wire that carries a
charge  per metre – spread evenly along the wire – something which
Coulomb’s Law could do, but would need a horrendous integral to do it.
This time, our surface is a cylinder, one metre long, with the wire running
down its axis. The cylinder has radius R. First, notice that the field lines
will run radially out from the wire. This has the consequence that the two
Page 92
Corrections March 2007
flat ends of the cylinder do not count in the integral. Think about this for
a moment, because this is important. The vector dS for these ends will
point normal to the surface – that means parallel to the wire. The vector
E, on the other hand points outward. Therefore E is perpendicular to dS,
and the dot product is zero.
Only the curved surface counts. Again, E will have the same magnitude
at all points on it because of symmetry, and E will be parallel to dS on
this surface. Once again, we have
 E  dS   E dS  E  dS  E 2RL
S
curved S
(5)
curved S
where L is the length of the cylinder, and hence 2RL is its curved
surface area. By Gauss’ Law, this must be equal to Q  0  L  0 , and
so
E

.
2 0 R
(6)
We notice that for a line charge, the “inverse square” of the Coulomb
Law has become an “inverse, not squared” law.
Before moving on, let us make two more points. Firstly, we have not
proved the equivalence of Gauss’ Law and Coulomb’s Law – we have
only shown that they agree in the case of calculating the field around a
fixed, point charge. However the two can be proved equivalent – but the
proof is a bit involved, and is best left to first (or second) year university
courses.
Secondly, let us think about a surface S which is entirely inside the same
piece of metal. E will be zero within a metal, because any non-zero E
(i.e. voltage difference) would cause a current to flow until the E were
zero. Therefore a surface entirely within metal can contain no total
charge!
Impossible, I hear you cry! Let us take a hollow metal sphere, with a
charge +Q at the centre of the cavity. How can the enclosed charge be
zero – surely it’s +Q! Oh, no it isn’t. Actually the total enclosed charge
is zero, and this enclosed charge is made up of +Q at the middle of the
cavity, and –Q induced on the inside wall of the hollow sphere! If the
sphere is electrically isolated, and began life uncharged, there must be a
+Q charge somewhere on the metal, and it sits on the outer surface of
the sphere.
6.1.2
Capacitors
A capacitor is a device that stores charge. To be more precise, a
capacitor consists of two conducting plates, with insulating space
Page 93
Corrections March 2007
between them. When the positive plate carries charge +Q, an equal
amount of negative charge is stored on the other. If certain insulating
materials are used to separate the plates (instead of air or vacuum), the
amount of charge that can be stored increases considerably. The
charge stored is proportional to the potential difference across it, and we
call the constant of proportionality the capacitance.
Gauss’ Law gives us a wonderful way to calculate the capacitance of
simple capacitors, and we will look at the calculation for a parallel plate
capacitor.
6.1.2.1 Parallel Plate Capacitor
At its simplest, a capacitor is shown in the figure below. The two plates
are square, and parallel. Each has area A and the distance between
them is denoted L. Let’s work out the capacitance. To do this, we first
suppose that there were a charge Q stored. In other words, there is a
charge –Q on the top plate, and +Q on the bottom plate. We can work
out the electric field in the gap using Gauss’ Law. We draw a
rectangular box-shape surface, with one of its faces parallel to, but
buried in the bottom plate, and the opposite face in the middle of the
gap.
Area A
Separation L
Voltage
difference V
When working out the surface integral
 E  dS
, only this face in the
S
middle of the gap counts. The face buried in the metal of the plate has
E=0, while the other four surfaces’ normals are perpendicular to the field.
Therefore
Q
0
  E  dS  AE .
(7)
S
We next work out the voltage. This is not hard, as by analogy from
gravitational work (chapter 1, equation 16)
Page 94
Corrections March 2007
Field = - d(Potential)/d(distance)
E = -dV/dx
(8)
Here, E is constant and uniform throughout the inter-plate gap, and so
V=Ex+c where c is a constant of integration. Thus the voltage difference
between the plates can be calculated; and from this the capacitance can
be worked out.
V  EL 
QL
0 A
 C
Q 0 A

L
V
(9)
We ought to give a word of caution at this point. In a real parallel plate
capacitor, the field near the edge of the plates will not point directly from
one plate to the other, but will ‘bow out’ a bit. Therefore the equation
given above is only true when these ‘edge effects’ are ignored. It turns
out that the equation is pretty good providing that L is much smaller than
both of the linear dimensions of the plates.
The equation also allows you to see the effects of wiring capacitors in
series or parallel. When two identical capacitors, each of capacitance C,
are connected together in parallel, the overall area A is doubled, so the
capacitance of the whole arrangement is 2C. On the other hand if the
capacitors are connected in series, the result is one capacitor with twice
the gap thickness L. Therefore the overall capacitance is C/2.
6.1.2.2 Decay of Charge on Capacitor
We next come to the case where a capacitor is charged to voltage V
(that is, a voltage V across the plates), and then connected in a simple
circuit with a resistor R. How long will it take to discharge?
To work this out, we need to use our characteristic equations for
capacitor and resistor. For the capacitor V=Q/C, for the resistor V=IR.
To solve the circuit we need to clarify the relationship between Q and I.
I
+Q
-Q
Here we need to take care. Depending on how the circuit has been
drawn, either I=dQ/dt or I=-dQ/dt. That is why it is essential that you
Page 95
Corrections March 2007
include in your circuit diagrams arrows to show the direction of current
flow for I>0, and which plate of the capacitor is the +ve one. Here a
positive I will discharge the capacitor, so I=dQ/dt.
The voltage across the capacitor is the same as that across the resistor,
so
dQ Q

dt C
.
dQ
Q

dt
RC
V  IR   R
(10)
This differential equation can be solved with an exponential solution:
Q(t )  Q0 e t RC
(11)
where Q0 was the initial charge on the capacitor after it was charged up.
Given that the voltage across the capacitor V(t)=Q(t)/C, the time
dependence of the voltage obeys a similar equation. The constant RC is
known as the time constant, and it is the time taken for the voltage (or
charge) to fall by a factor e (approx 2.7).
6.1.2.3 Energy considerations
Next, we need to know how much energy has been stored in a capacitor,
if its voltage is V and its capacitance C. The energy is actually ‘stored’ in
the electric field between the plates – but more of this later.
To work the energy out, we charge a capacitor up from scratch (initial
charge = 0), and continually measure the current flowing, and the
voltage across it. The energy stored must be given by
U   P dt   VI dt   V
dQ
d (CV )
dt   V
dt
dt
dt

 
dV
  VC
dt   CVdV  C  VdV  C 12 V 2 
dt
1
2
CV
2

(12)
Given that the energy stored must be zero when V=0, and there is no
electric field in the gap, this fixes the constant of integration as zero, and
we obtain the fact that energy stored = half the capacitance × the square
of the voltage across the gap. You could equally well say that the
energy is given by half the charge multiplied by the voltage.
Before we leave this formula, let us do some conjuring tricks with this
energy, assuming that the capacitor is a simple parallel plate device:
U  12 CV 2 
 0 AV 2
2L

 0 AEL 2
2L
Page 96
 12  0 E 2  AL
(13)
Corrections March 2007
thus the energy stored per unit volume of gap is 12  0 E 2 . Although we
have only shown this to be true for a perfect parallel plate capacitor, it is
possible to make any electric field look like rows and columns of parallel
plate capacitors arranged like a mosaic, and from this the proof can be
generalized to all electric fields.
6.1.2.4 Polarization
When we introduced capacitors, we mentioned that the capacitance can
be raised by inserting insulating stuff into the gap. How does this work?
Look at the diagram below. The stuff in the middle contains atoms,
which have positive and
negative
charges
within
them.
+
When
the
plates
are
+
charged, as shown, this pulls
+
the nuclei of the stuff to the
+
right, and the electrons of the
atoms to the left. The left
plate now has a blanket of negative charge, and the right plate has a
blanket of positive charge. This reduces the overall total charge on the
plates, and therefore reduces the voltage across the capacitor. Of
course the circuit can’t remove the “polarization” charges in the
‘blankets’ – as that would require the chemical breakdown of the
substance. So we have stored the same ‘circuit charge’ on the capacitor
for less voltage, and so the capacitance has gone up. The ratio by
which the capacitance increases is called the relative permittivity of the
substance (it used to be called the dielectric constant), and is given the
symbol r. In the presence of such a material, the 0 of all the equations
derived so far in this chapter needs to be multiplied by r.
+
-
6.2 Magnetism – when things move
So far, we have just considered electric charges at rest. Our next job is
surely to look at electric charges that have gone roaming, and then to
study magnetism – two things still to do? No. Actually we only have one
job, because magnetism is all about moving charges.
We ought to give one warning, though. Just because magnetism is
about moving charges, we can’t derive its formulae simply from
Coulomb’s Law and Classical mechanics.
We need Relativistic
mechanics! That is actually one route into relativity – it is the kind of
mechanics needed if electricity and magnetism are to be described
together. Put another way, your nearest piece of evidence for special
relativity is not in a particle accelerator or airborne atomic clock, but in
your wrist-watch (if it has hands), credit card, vacuum cleaner, fridge, CD
player, printer, hard disk drive, or wherever your nearest magnet is.
Page 97
Corrections March 2007
We shall demonstrate this at the end of the chapter. However, for the
moment, let us stick to what we need for the Olympiad, and let us carry
on calling it “magnetism” as opposed to “relativistic electricity”.
6.2.1
Magnetic Flux Density
If there is a magnetic field, there must be a measurement of the field
strength, and we call this the flux density, and give it the symbol B. The
field has a direction (from North to South), and is therefore a vector. The
fundamental fact of magnetism can be stated in two ways:
1.
If a wire of length L is carrying current I, and the wire is in a
magnetic field B, it will experience a force F, where F = L I × B. Written
without the vector cross product, this is F = B I L sin , where  is the
angle between the direction of the current, and the direction of the
magnetic field.
2.
If a charged particle, of charge Q is moving in a magnetic field B,
and it has velocity u, it will experience a force F, where F = Q u × B.
Written without the vector cross product, this is F = Q u B sin  where 
is the angle between the direction of motion, and the direction of the
magnetic field.
You can show that these descriptions are equivalent, by imagining the
wire containing N charges (each Q coulombs) per metre. If the wire has
length L, the total charge is QTOT=NQL, and this moves when the current
is flowing. If the current is I, this means that the charge passing a point
in one second is I, and hence I=NQu, where u is the speed. Therefore
F=LI×B
= L (NQu) × B
= (NQL) u × B
= QTOT u × B.
6.2.2
(14)
Doing the Corkscrew
Now that we have an expression for the force on a charge moving in a
magnetic field, we can work out the motion if a charge is thrown into the
vicinity of a magnet.
The most important fact is that the force is always at right angles to the
velocity. Therefore it never does any work at all, and it never changes
the kinetic energy (hence speed) of the object.
The next important fact is that if the velocity is parallel to the magnetic
field, there is no force – and the particle will just carry on going: as if the
magnet weren’t there at all.
Page 98
Corrections March 2007
This second fact is useful, because any velocity can be broken down (or
resolved) into two components – one parallel to the magnetic field (u
cos), and one perpendicular to it (u sin). The component parallel to
the field will be unchanged by the motion – it will stay the same, just as
Newton’s First Law requires.
We next need to calculate what the effect of the other component will be.
This will cause a force perpendicular to both the velocity and magnetic
field, and we already know from classical mechanics that when a force
consistently remains at right angles to the motion, a circular path is
obtained. We can calculate the radius from the equations of circular
motion:
Magnetic Force = Mass × Centripetal Acceleration
mu sin  
BQu sin  
R
2

R
mu sin 
BQ
(15)
Another useful measurement is the time taken for the particle to ‘go
round the loop’ once. Since its rotational speed is u sin, the time taken
to go round a 2R circumference is T  2R u sin  . We can also work
out the angular velocity:

u sin  BQ

R
m
(16)
The overall motion is therefore a helix, or corkscrew shape, with the axis
of the corkscrew parallel to the magnetic field. The radius of the helix is
given by R in equation (15), and the ‘pitch’ (that is, the distance between
successive revolutions), is equal to D = T u cos.
Helix, or corkscrew motion of an electron in a
magnetic field.
Please notice that all these formulae remain valid when the particle
starts going very quickly. The only correction that special relativity
requires is that we use the enhanced mass m   m rest . No further
Page 99
Corrections March 2007
correction is needed, because the speed remains constant, and
therefore  does not change.
6.2.3
Calculation of magnetic field strengths
So far, we have just thought about the effects of a magnetic field.
However before you can study what an electron will do in such a field,
you have to make the field! How do you do that?
At its simplest, magnetic fields are caused by electric currents. The
bigger the current: the bigger the field. These may be ‘real’ currents of
electrons in wires, or they can be the effective currents of electrons
‘orbiting’ their nuclei in atoms. The latter is responsible for the
permanent magnetic property of iron (and some other metals) – however
the process is quite involved and needs no further consideration for the
Olympiad.
We do need to worry about the magnetism caused by wires, however,
although only a brief description is necessary. The Olympiad syllabus
precludes questions that involve large amounts of calculus (hence
integration), and most field calculations require an integral. Therefore if
you need to know how big a magnetic field is, you will probably be given
the equation you need.
Nevertheless it will help to see how the calculations are done. There is a
method akin to Coulomb’s Law, and an alternative called Ampere’s Law.
We shall introduce both.
6.2.3.1 Magnetic Coulomb – The Biot Savart Law
To use this method, the wire (carrying current I) is broken down into
small lengths (dL), linked head-to-tail. To work out the size of the
magnetic field at point r, we sum the effects of all the current elements.
Let us take one current element at point s. Its contribution to the B field
at r is:
d Br  
 0 I  r  s 
dL
4 r  s 3
(17)
where (r-s) is the vector that points from the bit of wire to the point at
which we are calculating B. To work out the total field B(r), we integrate
the expression along the wire. In most cases, this can be put more
simply as
dB 
 0 I sin 
dL
4 d 2
Page 100
(18)
Corrections March 2007
where d is the distance from wire element to point of measurement, and
 is the angle between the current flow and the vector from wire to
measurement point.
It is possible to use this expression to calculate the magnetic field B on
the axis of a coil of wire with N turns, radius R, carrying current I, if we
measure the field at a distance D from the central point:
B

 0 INR 2
2 R2  D2

32
(19)
6.2.3.2 The Ampere Law
The alternative method of calculating fields is called the Ampere Law.
You remember that Gauss’ electrostatic law involved integrating a dot
product over a surface? Well, the Ampere Law involves integrating a dot
product along a line:
 B  dL  
0
I
(20)
L
In other words, if we choose a loop path, and integrate the magnetic flux
density around it, we will find out the current enclosed by that loop.
Let us give an example. This formula is very useful for calculating the
magnetic field in the vicinity of a long straight wire carrying current I.
Suppose we take path L to be a circle of radius R, with the wire at its
centre, and with the wire perpendicular to the plane of the circle, as
shown in the diagram.
Current I
Integration path round
the wire, keeping
constant distance from it.
R
We know that B points round the wire, and therefore that B is parallel
with dL (dL being the vector length of a small part of the path).
Page 101
Corrections March 2007
Furthermore, B=|B| must be the same all round the path by symmetry.
Therefore:
 B  dL   B dL  B  dL  2RB  
L
L
0
I
L
(21)
 I
B 0
2R
Another splendid use of the Ampere Law is in calculating the magnetic
field within the middle of a long solenoid with n turns per metre. This
time we use a rectangular path, as shown in the diagram.
Integration path
B-field direction
Only one of the sides of the rectangle “counts” in the integration – the
one completely inside. Of the other three, one is so far away from the
coil that there is no magnetic field, and the other two are perpendicular
to the B-field lines, so the dot product is zero. The rectangle encloses
nL turns, and hence a current of nLI. Therefore, Ampere’s Law tells us:
BL   0 nLI
(22)
B   0 nI
Before leaving Ampere’s Law, we ought to give a word of caution. The
Ampere Law is actually a simplified form of an equation called the
Ampere-Maxwell Law 22 ; and the simplification is only valid if there are no
22
For the curious, the full Ampere-Maxwell Law states

L
1
0
  dS
B  dL  I    0  rel E
S
where the surface S is any surface that has as its edge the loop L. This reduces to equation
(20) when E is constant.
Page 102
Corrections March 2007
changing electric fields in the vicinity. Therefore you would be on dodgy
ground using Ampere’s Law near a capacitor that is charging up!
6.2.4
Flux, Inductance & Inductors
To sum up the last section: if there is a current, there will be a magnetic
field. Furthermore, the strength of the magnetic field is proportional to
the size of the current. It turns out, however, to be more useful to speak
of the magnetic flux . This is the product of the field strength B and the
cross sectional area of the region enclosed by the magnetic field lines.
We visualize this as the total ‘number of field lines’ made by the magnet.
We then write
  LI .
(23)
The total amount of magnetic field (the flux) is proportional to the current,
and we call the constant of proportionality L – the self-inductance (or
inductance for short). Any coil (or wire for that matter) will have an
inductance, and this gives you an idea of how much magnetic field it will
make when a current passes. You might think of an analogy with
capacitors – the capacitance gives a measure of how much electric field
a certain charge will cause (since C-1 = V/Q and V is proportional to E).
Now, this magnetic field is important, because a changing current will
cause a changing magnetic field, and this will generate (or induce) a
voltage, and therefore upset the circuit it is in. This is something we
need to understand better – but before we do so, let us remind ourselves
of the laws of electromagnetic induction:
6.2.5
Generators & Induction
If a wire ‘thinks’ it is moving magnetically, a voltage is induced in it. It
doesn’t matter whether the wire is still, and the magnetic field is moving
or changing; or whether the wire is moving and the magnetic field is still.
However for the two situations, different equations are used. The two
equations are equivalent – however it is easier to remember them both
than to prove the equivalence.
6.2.5.1
Stationary field: Moving wire
motion of wire
electrons in wire pushed to right
Magnetic field B down into paper.
Page 103
Corrections March 2007
Here the voltage is easy to calculate. The electrons in the wire must
move with the wire, and if the wire is in a magnetic field, they will
experience a force pushing them along the wire. This causes them to
bunch up at one end of the wire – which in turn sets up an electric field
which discourages further electrons to join the party. Assuming that the
wire has length L, and is being moved at speed u through a uniform
magnetic field B (perpendicular to u and L – if B is not so inclined, take
only the component of B which is), once equilibrium is established
Electric force balances magnetic force
qE  quB
V
 uB
L
V  LuB
(24)
6.2.5.2 Stationary wire: Changing field
Here the voltage induced across the ends of a circular loop of wire
(complete circuit, apart from the small gap) is given by
V
d
dt
 B  dS 
S
d
dt
(25)
6.2.5.3 Equivalence?
The two expressions are very similar, as we shall see when we look at
the first from a different perspective. Look at the diagram – we have
completed a loop by using a very long wire.
Region of magnetic
field B, pointing
down into paper
End wire moved in
direction of white
arrow. Distance
moved equals ut,
where u is speed.
L
In time t, the area is reduced by
Lut , so the magnetic flux
enclosed by the wire is reduc ed
by BLut . The rate of change of
flux = induced voltage = BLu.
Page 104
Corrections March 2007
The total flux enclosed in the loop is equal to B × Area. In one second,
the Area changes (decreases) by Lu, and hence the flux changes by
BLu, and so we see that the first equation is a special case of the
second. The second is better as it also allows us to calculate what will
happen if the wire is stationary.
6.2.5.4 Direction of Induced Voltage
When equation (25) is written down, it is customary to put a minus sign
in front of the derivative. The significance of this negation was Lenz’s
discovery. When the voltage is induced in a complete circuit, it will try to
(and succeed in) driving a current. This current will produce a magnetic
field. Lenz postulated that this ‘produced’ magnetic field always
opposed the change being made.
Let us have an example. Imagine a large coil of wire (say, in a motor),
with a decent sized current flowing in it. Now let us try and lower the
current by reducing the voltage of the supply. This causes a reduction in
the magnetic field, which in turn induces a voltage in the wire, which
pushes a current in a desperate attempt to keep the original current
going. On the other hand, if I were to try an increase the current in the
motor (by increasing the supply voltage), the opposite would happen: the
greater current causes the magnetic field to grow, which induces a
voltage, which pushes a current to oppose this increase.
This is the origin of the phrase “back emf” to refer to the voltage induced
across an inductor.
Now for a word about the minus sign. Yes, the voltage does go in
opposition to the change in current, so I suppose one ought to write
equation (25) with a minus sign. However if you do, please also write
Ohm’s law as V = IR, since the voltage opposes the current in a
resistor. I would prefer it, however, if you used common sense in
applying your notation and were not stuck in ruts of “always” or “never”
using the minus sign. We all remember which way V and I go in a
resistor without being nagged about conventions, so I hope that there is
no need for me to nag you when inductors come on the scene.
6.2.6
Inductors in circuits
Just as a capacitor requires an energy flow to change the voltage across
it, an inductor requires an energy flow to change the current through it. It
doesn’t give in without a fight.
Let me illustrate this with a demonstration – or at least the story of one.
A nasty physics teacher (yes, they do exist...) asked a pupil to come and
hold one wire in his left hand, and one in his right – completing the circuit
with his body. He grasped the first wire, and then the second – steeling
himself for the shock which never came. The teacher then stalled him
with questions, and kept him there, while he surreptitiously, slowly
increased the current in the circuit, which also included an inductor.
Page 105
Corrections March 2007
Finally, the teacher said, “OK, now you can go back to your seat.” The
unsuspecting pupil let go of the wires suddenly, and was very surprised
when the inductor – indignant to have its current shut off so quickly –
made its displeasure known with an arc from the wire to the pupil. It is
foolish in the extreme to starve an inductor of its current. Its revenge will
be short, but not sweet.
To see this, let us combine equations (23) and (25)
V
dI
d d
 LI  L
dt
dt dt
(26)
Equation (26) is the definitive equation for inductors, just like Q=CV was
for capacitors, and V=IR is for a resistor.
Again, we need to bear in mind the comments above, that the voltage is
in the direction needed to oppose the change in current. You will often
see (26) with a minus sign in it for that reason.
Let us now calculate how much energy is stored in the device (actually in
its magnetic field)
U   P dt   VI dt   L
dI
I dt  L  I dI 
dt

1
2
LI 2

(27)
Since it seems sensible for the device to hold no energy when there is
no current and no field, we take the constant of integration to be zero.
6.2.7
Relativity and Magnetism
At the beginning of the chapter we stated boldly that magnetism could be
derived entirely from electrostatics and special relativity. Now is the time
to justify this. We shall do so by deriving the same result two ways –
once using magnetism, and once using relativity.
The phenomenon we choose is the mutual attraction of two parallel
wires carrying equal current in the same direction, and we shall calculate
the attractive force per metre of wire.
Page 106
Corrections March 2007
Both wires carry
current I
Force on right
hand wire
Direction of Bfield due to left
hand wire
R
6.2.7.1 A Classical Magnetic calculation
Equation (21) tells us that the magnetic field at a distance R from a
straight infinite wire is
B
0 I
2R
Therefore, the attractive force experienced by one metre of parallel
conductor also carrying current I is
 I2
F
 IB  0
2R
L
(28)
6.2.7.2 A Relativistic calculation
Each wire contains positive ion cores (say Cu+ for a copper wire), and
free electrons. Let us imagine the situation in the diagram below, with
conventional current flowing downwards in both wires. The ion cores are
stationary, while the electrons move upwards. If the free (electron)
charge per metre of cable is called 0, then the current is related to the
electron speed u by the equation I=0u.
Page 107
Corrections March 2007
Distance
between
ion cores
From perspective of
ions in second wire,
first wire appears
negati vely charged.
Wires attract.
Distance
between
electrons
appears
contracted
Let’s imagine the situation as perceived by an ion core in the second
(right hand) wire. It sees the ion cores in the other wire stationary, with
charge density 0 and finds them repulsive. However it also sees the
electrons on the other wire, and is attracted by them. The electrons are
travelling, and therefore our observant ion core sees the length between
adjacent electrons contracted. Therefore as far as it is concerned, the
electron charge density is higher than the ion core charge density by
factor   1  u 2 c 2  . Therefore its overall impression is attraction –
with a total effective electric field (as derived in equation 6)
1 2
Total field = Field due to ion cores + Field due to loose electrons
0
 0

2 0 R 2 0 R
0
   1
2 0 R
E
 u2
  2
 2c
(29)
 0

 2 0 R
where the final stage has made use of a Binomial expansion of (-1) to
first order in u/c. Now the total charge of ion cores experiencing this field
per metre is of course 0. Therefore the total attraction of the ion cores
in the second wire to the first wire (electrons & ion cores) is
 u2
Fions  QE   2
 2c
2
I
 2
4c  0 R
  20

 2 0 R
Page 108
(30)
Corrections March 2007
This is the force experienced by the ion cores in the right hand wire. By
an exactly equivalent argument, the electrons in the right hand wire see
their counterparts in the left wire as stationary, and see the left hand ion
cores bunched up, and therefore more attractive. Therefore the electrons
in the right hand wire experience an equal attraction, and the total
attractive force between the wires is twice the figure in equation (30).
Finally, if you get a book of physical constants and a calculator, you will
discover that  0  0  c 2 . Therefore the total force agrees exactly with
our magnetic calculation in equation (28).
6.3 Circuits – putting it together
In this section, we look at combining resistors, capacitors and inductors
in electrical circuits. There are two reasons for doing this. Firstly, once
you have left school, you will be faced with complicated electronic
networks, and you need to be able to analyse these just as well as the
simple series and parallel arrangements you dealt with in the classroom.
Secondly, engineers frequently use electric circuits as models or
analogies for other systems (say, an oscillating bridge or the control of
the nervous system over the muscles in a leg) – the better you
understand electric circuits, the better you will understand any linked
system.
6.3.1
Circuit Analysis
Our aim here is to be able to solve a circuit like the one below. The
circles represent constant-voltage sources (a bit like cells or batteries)
and the linked circles represent constant-current sources. Our aim is to
find voltage difference across each component, and also to work out the
current in each resistor.
A
R1
V1
B
R2
C
R3
I1
R4
D
V2
E
In order to solve the circuit, we use two rules – the Kirchoff Laws.
Kirchoff’s 1st says that the total current going into a junction is equal to
the total current leaving it. Therefore, at B in the circuit below, we would
say that IBE = IAB + ICB, where IBE means the current flowing from B to E
(through R4).
Kirchoff’s 2nd Law is that voltages always add up correctly. In other
words, no matter which route we took from E to B, say, we would agree
on the voltage difference between E and B. In symbols, if VBE means
Page 109
Corrections March 2007
the difference in potential (as measured by a voltmeter) between B and
E, then we have VBE = VAE + VBA. This is basically the same thing as the
law of conservation of energy. The voltage (or, more strictly, the
potential) at B, VB, is the energy content of one coulomb of charge at B.
In travelling to E, it will lose VB-VE joules, irrespective of the route
taken. 23 In fact, we assume the truth of Kirchoff’s 2nd Law whenever we
say, “let’s call the voltage at A ‘VA’,” for we are assuming that the voltage
of A does not depend on the route used to measure it.
Using these two rules, and the equation for the current through a resistor
(for example, VBA = IBA R1), we may write down a set of equations for the
circuit. Notice that because currents are said to go from + to , this
means that if VAB (the voltage of A, measured relative to B) is positive,
then VA is bigger than VB, and hence IAB will be positive too. To make
the notation easier we will take the potential at E to be zero. In symbolic
form, this means that we shall call VBE (that is, VBVE) VB for short.
Kirchoff’s First Law:
IEA = IAB;
IBE = IAB+ICB;
I1 = ICB+ICD;
ICD = IDE
Kirchoff’s Second Law:
VB = IBE R4
= VA +VBA = V1  IAB R1
= VD + VCD + VBC = V2 + ICD R3  ICB R2
After elimination, the equations reduce to two:
V1 – IAB R1 = V2 + I1 R3 – (R3 + R2) ICB = (IAB + ICB) R4,
and from these the currents IAB and ICB can be found (after a bit of messy
algebra).
After this, the remaining currents and voltages are
straightforward to determine.
These principles can be used to solve any circuit. However, as networks
get bigger, it is useful to find more prescriptive methods of solution,
which could be used by a computer. We shall cover two methods here –
for certain problems, they may be more efficient than the direct
application of Kirchoff’s Laws.
23
To see why the Law of Conservation of Energy is involved, let us suppose that our coulomb
of charge would lose 5J going from B to E via A, whereas it would lose 3J in going directly.
All it would have to do is go direct from B to E, then back to B via A and it would be back
where it started, having gained 2J of energy! This is not allowed.
Page 110
Corrections March 2007
6.3.1.1 Method of Superposition
The method of superposition relies on the fact that for a simple resistor,
the current is proportional to the voltage. It follows that if current I1
causes a voltage difference of 3V, and current I2 causes a voltage
difference of 5V, then current I1+I2 will cause an 8V p.d. across the
component. Here is the procedure:






Choose one of the supply components.
Remove the other supply components from the circuit. Replace voltage
sources with direct connections (short circuits), and leave breaks in the
circuit where the current sources were (open circuits).
Calculate the current in each wire, and the voltage across each
component.
Repeat the procedure for each supply component in turn.
The current in each wire for the original (whole) circuit is equal to the
sum of the currents in that wire due to each supply unit.
The voltage across each component in the original (whole) circuit is
equal to the sum of the voltages across that component due to each
supply unit.
Let’s use this method to analyse the circuit above. We start by
considering only source V1. Removing the other supply components
gives us a circuit like this.
A
R1
B
R2
C
R3
D
R4
V1
E
This circuit is easier to analyse as it only has one supply. Supply V1
feeds a circuit with resistance
R1 + {R4 // (R2+R3)}
 R1 
R4 R2  R3 
R4  R2  R3
where // means ‘in parallel with.’ Accordingly, the current supplied by V1
(and the current through R1 which is in series with it) is equal to V1
divided by this resistance. The voltages of points B, C and D can be
calculated, as can the current in each wire. We make a note of the
values, and add to them the results of analyses of circuits only
containing I1 and only containing V2.
Page 111
Corrections March 2007
You may find this method good in the sense that you only have to deal
with one supply component at a time – and therefore all you need to
know is how to combine resistors (something you’ve done before).
Having said that, we end up analysing three circuits rather than one, so it
is more time consuming.
Before leaving the method, you may be curious why voltage sources
were replaced with short circuits, and current sources with open circuits.
Here’s the reason. A voltage source does not change the voltage across
its terminals, no matter what the current is (d Voltage / d Current =
Requivalent = 0). The only type of resistor which behaves likewise is a
perfect conductor (0). Similarly, a current source does not change its
current, no matter what the voltage (d Current / d Voltage = 1 / Requivalent
= 0). The equivalent resistor in this case is a perfect insulator (∞)
which lets no current through ever.
6.3.1.2 Method of Loop Currents
Here we break the circuit down into the smallest loops it contains. Here
there are three loops:
A
R1
V1
IL1
B
C
R2
R4
IL2
R3
I1
IL3
D
V2
E
 E to A to B and back to E (loop 1),
 E to B to C to E (loop 2), and
 E to C to D to E (loop 3).
We call the current in loop 1 “loop current” number one (IL1), with IL2 and
IL3 representing the currents in the other two loops. We then express all
other currents in terms of the loop currents. Clearly, IAB = IL1, since R1 is
in the first loop alone. Similarly, IBC = IL2, and ICD = IL3.
The current through R4 is more complex, since this resistor is part of two
of the loops. We write IBE = IL1  IL2. Here IL1 is positive, since IBE is in
the same direction as IL1, whereas IL2 (which goes from E to B then on to
C) is in the opposite direction. These designations automatically take
care of Kirchoff’s First Law. Notice that by this method, I1 = IL3  IL2.
Each loop now contributes one equation  Kirchoff’s 2nd law around that
loop. Clearly, if you go all the way round the loop, you must return to the
voltage you started with. Taking the first loop as an example, we have:
0
= VAE + VBA + VEB
Page 112
Corrections March 2007
= V1  IAB R1  IBE R4
= V1  IL1 R1 + (IL1  IL2) R4.
In a similar way, we write equations for each of the other two loops 24 .
We then have three equations in three unknowns (the three loop
currents), which can be solved. The end result is the same as for a
direct ‘sledgehammer’ approach with Kirchoff’s Laws – but the method is
more organized.
6.3.2
Alternating Current
Having looked at circuits with resistors in them, we next turn our
attention to circuits with inductors and capacitors as well. For a direct
current, the situation is easy. After a brief period of settling down, there
is no voltage drop across an inductor (because the current isn’t
changing), and a capacitor doesn’t conduct at all.
For alternating currents the situation is more complicated. Let us
suppose that the supply voltage is given by V=V0 cos t. It turns out that
the circuit will settle down to a steady behaviour (called the steady
state). Once this has happened, the voltage across each component
(and the current through each component) will also be a cosine wave
with frequency , however it may not be in phase with the original V.
6.3.2.1 Resistor, capacitor and inductor
We start with the three simplest circuits – the lone resistor, the lone
capacitor and the lone inductor, each supplied with a voltage V=V0 cos
t.
For the resistor, I=V/R, so the result is straightforward.
For the capacitor, Q=VC, and if we take I as positive in the direction
which charges the capacitor, then
dQ d
 CV0 cos t
dt dt
  CV0 sin t
I
  CV0 cost   
(31)
1
2
  CV0 cost  12  
For the inductor, V  L dI dt , so
24
It may help when writing the equations to notice the pattern: voltage sources count
positively if you go through them from  to +, but negatively if you go from + to . The
voltages across resistors (e.g. VBA = I R1) count negatively if you go through them in the same
direction as the current, and positively if you go through them the opposite way to the current.
Page 113
Corrections March 2007
dI V0
 cos t
dt L
V
I  0 sin t
L
V
 0 cost  12  
L
(32)
where we have taken the constant of integration to be zero. Failure to
do so would lead to a non-zero mean current, which is clearly impossible
as the mean supply voltage is zero.
6.3.2.2 Reactance and Impedance
For resistors, the current and voltage are proportional, and consequently
are in phase – one peaks at the same time as the other. For the other
two components, this is not the case. The voltage is /2 radians (or 90°)
out of phase with respect to the current. Inductor currents peak 90° later
than the voltage (the current lags the voltage), whereas capacitor
currents peak 90° before the voltage (the current leads the voltage).
Nevertheless, the amplitude of the voltage is still proportional to the
amplitude of the current, and we call the ratio of the amplitudes the
reactance (X).
X L  L
XC  
1
C
(33)
By convention, we take reactance to be positive if the current lags the
voltage by 90°, and negative if it leads by 90°. For capacitors and
inductors in series, the total reactance is equal to the sum of the
individual components’ reactances – just as resistances add in series.
Similarly, the formula for combining reactances in parallel is the same for
that used for the resistance of resistors wired in parallel.
When a circuit is constructed with resistors, capacitors and inductors,
then we need a way of analysing a circuit with both resistances and
reactances. We visualise the situation using a 2D (phasor) diagram.
For any component or circuit, both voltage and current are represented
by vectors. The length of the lines gives the amplitude, and the angle
between the vectors gives the phase difference. By convention, we
imagine the vectors to rotate about the origin in an anticlockwise
direction (once per time period of the alternating current). The vectors
for a resistor, capacitor and inductor are shown below.
Page 114
Corrections March 2007
V=I XL
Origin
Origin
I
V=IR
Origin
I
I
V=I XC
Resistor
Inductor
Capacitor
As the arrows rotate anticlockwise, for the inductor, V comes before I.
With the capacitor, I comes before V. This accurately represents the
phase relationships between voltage and current for these components.
For a set of components in series, the current I will be the same for all of
them. We usually draw the current pointing to the right. Voltages across
inductors will then point up, those across resistors point right, and those
across capacitors point down. By adding these voltages vectorially, we
arrive at the voltage across the set of components – and can calculate
its amplitude and phase relationship with respect to the current.
Similarly, for components in parallel, the voltage will be the same for
each. We thus put voltage pointing to the right. Currents in capacitors
now point up, currents in resistors point right, and currents in inductors
point down. The total current is given by the vector sum of the individual
currents.
In all cases, we call the ratio of the voltage amplitude to the current
amplitude the impedance (Z) irrespective of the phase difference
between the current and voltage. 25 In general the impedance of a
component is related to resistance and reactance by Z2 = R2 + X2.
6.3.2.3 Complex Numbers and Impedance
If you are familiar with complex numbers, there is an easier way of
describing all of this, using the Argand diagram in place of 2-dimensional
vectors. The impedance Z is a now a complex number Z = R + iX, with
R as its real part and X as its imaginary part.
The complex impedances of a resistor, capacitor and inductor are
accordingly written as R, i/C and iL respectively. The impedance of
a set of components in series is given by the sum of the individual
impedances. For a parallel network, Z1 of the network is given by the
25
In other words, a resistance is a special kind of impedance with zero phase difference, and
a reactance is a special kind of impedance when the phase difference is 90°.
Page 115
Corrections March 2007
sum of Z1 for each component, where ‘inverse’ (or ‘reciprocal’) is
calculated in the usual way for complex numbers.
6.3.2.4 Root Mean Square values
You will also need to remember the definition of RMS voltage and
current in an a.c. circuit. For a resistor, remember that the RMS supply
voltage is the d.c. voltage which would supply the same mean power to
the device.
P
6.3.3
2
Vrms
V 2 cos 2 t V02
 0

R
R
2R
1
V0
 Vrms 
2
(34)
Resonance
One further circuit needs a mention, and that is the simple circuit of an
inductor and a capacitor connected together, as shown in the diagram
below. Both the voltage and current for the two components must be the
same, and so with the sign conventions chosen in the diagram:
Direction of positive current I
C
+Q
L
-Q
Q
d 2Q
dI
L
 L 2
C
dt
dt
2
d Q
1

Q
2
LC
dt
(35)
This is an equation of ‘simple harmonic motion’ with angular frequency
, where  2  1 LC .
This circuit can therefore oscillate at this
frequency, and this makes it useful in radio receivers for selecting the
frequency (and hence radio station) which the listener wants to detect.
6.4 Questions
1. Calculate the size of the repulsion force between two electrons 0.1nm
apart.
Page 116
Corrections March 2007
2. In this question, you will make an estimate for the size of a hydrogen atom.
Suppose an electron moves in a circular path around the proton, with
radius r. Calculate, in terms of r, the potential energy of the atom (it will be
negative, of course), the speed of the electron, and its kinetic energy.
Now write down an expression for the total energy of the electron. Find
the value of r which minimizes this total energy, and compare it to the
measured radius of a hydrogen atom, which is about 5×10–11 m.
3. What fraction of the electrons in the solar system would have to be
removed in order for the gravitational attractions to be completely
cancelled out by the electrostatic repulsion?
4. A cloud of electrons is accelerated through a 20kV potential difference (so
that their kinetic energy of each coulomb of electrons is 20kJ). Calculate
their speed.
5. A beam of 20kV electrons is travelling horizontally. An experimenter
wishes to bend their path to make them travel vertically (at the same
speed) using a region with a uniform electric field. This region is square
with side length 5cm. Calculate the size and direction of electric field
needed to do this. What would happen to a beam of 21kV electrons
passing this region?
6. A different experimenter wishes to bend the beam of 20kV electrons using
a magnetic field. She chooses to bend the beam round a circular path of
radius 3cm. What magnitude and direction of magnetic field is needed?
What would happen to a beam of 21kV electrons passing this region?
7. I wish to make a 1T magnetic flux density inside a long coil (or solenoid)
with radius 5mm. I use wire which can carry a current of 4A. How many
turns per metre of coil are needed?
8. ‘Clamp’ ammeters used by electricians can measure the current in a wire
without needing to break the wire. A metal loop encloses the wire, and the
magnetic field around the wire is measured. If the loop is circular, with
radius 3cm, and is centred on the wire, calculate the magnetic flux density
measured when the current in the wire is 100A.
9. Calculate the impedance of a 20 resistor wired in series with a 3mH
inductor when fed with alternating current of 50Hz. A capacitor wired in
parallel with this combination causes the overall reactance of the circuit to
become zero at 50Hz (in other words, the voltage is in phase with the total
current). Calculate the capacitance of the capacitor.
Page 117
Corrections March 2007
Finally, to convert from kilograms to joules, multiply by c2 (the speed of
light squared).
A particular case of a nuclear reaction is the ‘annihilation’ reaction, in
which a particle and its antiparticle (say an electron and a positron) react
together. The matter vanishes, and the energy appears in the form of
two gamma rays. 28
7.5 Questions
1. Calculate the wavelength and frequency of the quantum associated with a
60g ball travelling at 40m/s. Why don’t we observe interference effects
with balls such as this?
2. Blue light has a wavelength of approximately 400nm, while red light has a
wavelength of approximately 650nm. Calculate the energies of photons of
blue and red light (a) in joules (b) in electron-volts (eV). One electron-volt
is equal to 1.602×10–19 J.
3. Work out the wavelengths of light emitted when electrons from the n=5, 4,
and 3 levels ‘descend’ to the n=2 level. Why do you think that these
transitions were more important in the historical development of atomic
theory than the ‘more fundamental’ transitions going down to the n=1
level?
4. Calculate the energies of the n=1, 2, 3, 4 and 10 levels for an ionized
helium atom (a helium nucleus with a single electron).
5. Calculate the size of a muonic hydrogen atom in comparison with a normal
hydrogen atom. A muonic hydrogen atom has a muon rather than an
electron moving near a proton. The muon has a charge equal to that of an
electron, but its mass is 207 times greater.
6. In this question, you will make an estimate for the size of a hydrogen atom.
Suppose that the atom’s radius is r. Then the uncertainty in the electron’s
position is 2r. Use the uncertainty principle to work out the uncertainty in
its momentum, and from this work out its typical kinetic energy, in terms of
r. The electron’s typical electrostatic energy is –e2/40r2. Find the value
of r which minimizes the total energy of the electron.
7. Calculate the energy liberated in the fusion reaction 21 H  31 H 42 He 01 n .
The masses of the particles are given in the table in unified mass units (u).
1u = 1.660431027 kg.
28
When analysing the collision, you find that you can not satisfy momentum and energy
conservation at the same time if only one photon is produced.
Page 124
Corrections March 2007
7 Small Physics
The rules, or laws, of classical mechanics break down typically in three
cases. We have seen that when things start going quickly, we need to
take special relativity into account. Another form of relativity – the
general theory – is needed when things get very heavy, and the
gravitational fields are strong. The third exception is very mysterious –
and occurs often when we deal with very small objects like atoms and
electrons. This is the realm of quantum physics, and many of its
discoverers expressed horror or puzzlement at its conclusions and
philosophy.
Having said that, there is no need to be frightened. While there is much
we do not understand, a set of principles have been set up which allow
us to perform accurate calculations. Furthermore, those calculations
agree with experiment to a high degree of accuracy. The development
of the transistor, hospital scanner, and many other useful devices testify
to this. The situation is analogous to a lion-tamer who can get the lion to
jump through a hoop, though she doesn’t know what is going on inside
the lion.
7.1 Waves and Particles
Quantum objects, like electrons and photons (packets of light) are
difficult to describe. As physicists, we have two models, or descriptions,
which we are comfortable using – the wave and the particle.
Waves can interfere, they have a wavelength, frequency and intensity,
and they carry energy by means of fluctuations in a medium. The
intensity is continuous – it can take any value.
Particles on the other hand, are lumps. They possess individual
masses, energies and momenta. They most certainly do not interfere –
if you add 1 apple to 1 apple, you always get 2 apples. Finally they only
come in integer numbers. You can have one, or two, or 45 678 543; but
you can‘t have half.
The electron fits neither description. Light fits neither description. The
descriptions are too simplistic. However there are instances when the
particle description fits well – but it doesn’t always fit. There are also
instances when the wave description fits well – but it doesn’t always fit.
Given that a particular electron beam may behave like particles one
minute, and waves the next, we need some kind of ‘phrase book’ to
convert equivalent measurements from one description to the other.
Quantum theory maintains that such a ‘phrase book’ exists.
Page 118
Corrections March 2007
The total number of particles (in the particle picture) is related to the
intensity (in the wave picture). The exact conversion rate can be
determined using the principle of conservation of energy.
The energy per particle (in the particle picture) is related to the frequency
(in the wave picture) by the relationship
Energy of one particle (in J) = h × Frequency of wave (Hz),
(1)
where h is the Planck constant, and has a value of 6.63×10-34 Js. You
may also come across the constant ‘h-bar’   h 2 , which can be used
in place of h if you wish to express your frequency as an angular
frequency in radians per second.
The momentum per particle (in the particle picture) is related to the
wavelength of the wave (in the wave picture) by the relationship
Momentum of one particle (kg m/s) 
h
.
Wavelength (m)
(2)
7.2 Uncertainty
The bridge between wave and particle causes interesting conclusions.
We have seen in the chapter on Waves that a wave can have a welldefined frequency or duration (in time), but not both. This was
expressed in the bandwidth theorem:
f t  1 .
When combined with our wave-particle translation, we obtain a
relationship between energy and time:
E t  h .
(3)
In other words, only something that lasts a long time can have a very
well known energy.
Let us have an example. Suppose a nucleus is unstable (radioactive),
with a half-life T. Seeing as the emission of the radiation is a process
that typically ‘takes’ a time T, the energy of the alpha particle (or
whatever) has an inherent uncertainty of E  h T . If we were watching
a spectrometer, monitoring the radiation emitted, we would expect to see
a spread of energies showing this level of uncertainty.
The bandwidth theorem also has something to say about wavelength:
1
  x  1 .

Page 119
Corrections March 2007
This has the quantum consequence:
p x  h .
(4)
This is frequently stated as, “You can’t know both the momentum and
the position of a particle accurately.” It might be better stated as, “Since
it is a bit like a wave, it can not have both a well defined position and
momentum.”
We can use this to make an estimate for the speed of an electron in an
atom. Atoms have a size of about 10-10 m. Therefore, for an electron in
an atom, x  10 10 m . So, using equation (4), p  10 23 kg m/s . Given
the electron mass of about 10-30 kg, this gives us a speed of about
107 m/s – about a tenth the speed of light!
Caution: Please note that we haven’t defined precisely what we mean
by uncertainty (). That is why we have only been able to work with
approximate quantities. In more advanced work, the definition can be
tightened up (to mean, say, standard deviation). However it is better for
us to leave things as they are. In any case, it is never wise to state
uncertainties to more than one significant figure!
7.3 Atoms
Putting things classically for a moment, the electron orbits the nucleus.
While a quantum mechanic thinks this description very crude, we shall
use it as a starting point.
Now, let’s imagine the electron as a wave. For the sake of visualization,
think of it as a transverse wave on a string that goes round the nucleus,
at a distance R from it. If the electron wave is to make sense, the string
must join up to form a complete circle. Therefore the circumference
must contain a whole number of wavelengths.
2R  n
2R 
nh
p
nh
pR 
 n
2
L  n
(5)
The conclusion of this argument is that the angular momentum of the
electron, as it goes round the nucleus, must be in the ħ-times table.
The argument is simplistic, in that the quantum picture does not involve
a literal orbit. However, the quantum theory agrees with the reasoning
above in its prediction of the angular momentum.
Page 120
Corrections March 2007
Given that the angular momentum can only take certain values (we say
that it is quantized – it comes in lumps), we conclude that the electron
can only take certain energies. These are called the energy levels, and
we can work out the energies as follows: 26
L2
n2 2

2 I 2mR 2
,
Ze 2
Potential Energy  
4 0 R
Kinetic Energy 
(6)
where Z is the number of protons in the nucleus. We are, of course,
ignoring the other electrons in the atom – hence this model is only
directly applicable to hydrogen. 27 Next, we use the relationships derived
in section 1.2.2, where we showed that for a Coulomb attraction,
Potential energy = -2 × Kinetic Energy
Potential energy = 2E
Kinetic energy = -E
(7)
where E is the total energy of the electron. We may use this information
to eliminate the radius in equations (6), obtaining:
E
Z 2e4m
24 0  
2

1
.
n2
(8)
When dealing with atoms, the S.I. units can be frustrating. A more
convenient unit for atomic energies is the electron-volt. This is the
energy required to move an electron through a potential difference of
one volt, and as such it is equal to about 1.60×10-19 J. In these units,
equation (8) can be re-written:
Z2
E   2  13.6 eV .
n
(9)
This form should be remembered. It will help you to gain a ‘feel’ for the
energies an electron can have in an atom, and as a result, it will help you
spot errors more quickly.
26
The kinetic energy is calculated using the relationships derived in chapter 3. If you do not
wish to go in there, a simpler derivation can be employed. L=mvR, where v is the speed.
Therefore the kinetic energy mv2/2 = L2/(2mR2).
27
Hydrogen, that is, and hydrogen-like ions: which are atoms that have had all the electrons
removed apart from one.
Page 121
Corrections March 2007
When an electron moves from one level (n value) to another, energy is
either required or given out. This is usually in the form of a photon of
light that is absorbed or emitted. The energy of the photon is, as usual,
given by the Planck constant, multiplied by the frequency (in Hz) of the
light.
If an electron moves from orbit n1 to n2, where n2<n1, the frequency of
photon emitted is therefore given by:
 1
1 
f  Z 2  2  2 3.29  1015 Hz .
 n 2 n1 
(10)
Similarly, the formula gives the frequency of photon required to promote
an electron from n2 to n1. The frequency of photon required to remove
the electron completely from the atom (if it starts in level n2) is also given
by equation (10), if n1 is taken as infinite.
7.4 Little Nuts
As far as Romans were concerned, the stones in the middle of olives
were ‘little nuts’ or nuclei. We shall thus turn our attention to ‘nutty
physics’.
The nuclear topics required for the International Olympiad are common
to the A-level course. In this book we shall merely state what knowledge
is needed. You will be able to find out more from your school textbook.
7.4.1
Types of radiation
Alpha decay: in which a helium nucleus (two protons and two neutrons)
is ejected from the unstable nucleus.
Beta decay: in which some weird nucleonic processes go on. In all beta
decays, the total number of nucleons (sometimes called the mass
number) remains constant.
In - decay (the most common), a neutron turns into a proton and an
electron. The electron is ejected at speed from the nucleus.
There are two other forms of beta radiation. In + decay, a proton turns
into a neutron and an anti-electron (or positron). The positron flies out of
the nucleus, and annihilates the nearest electron it sees.
The
annihilation process produces two gamma rays.
The other permutation is electron capture () in which an electron is
captured from an inner (low n) orbit, and ‘reacts’ with a proton to make a
neutron. This phenomenon is detected when another electron descends
to fill the gap left by the captive – and gives out an X-ray photon as it
does so.
Page 122
Corrections March 2007
Gamma decay: in which the nucleus re-organizes itself more efficiently,
leading to a drop in its internal potential energy. This energy is released
as a burst of electromagnetic radiation – a gamma ray photon. By
convention high energy photons are called X-rays if they come from the
electrons in an atom, and gamma rays if they come from a nucleus.
7.4.2
Radioactive decay
It is beyond the wit of a scientist to predict when a particular nucleus will
decay. However we have so many radionuclides in a sample that the
average behaviour can be modelled well.
The rate of decay (number of decays per second) is proportional to the
number of nuclei remaining undecayed. This ‘rate of decay’ is called the
activity, and is measured in Becquerels (Bq). We define a parameter 
to be the constant of proportionality:
dN
 N
dt
N  N 0 e  t
I 
(11)
I  N 0 e  t  I 0 e  t
where N0 is the initial number of radionuclides, and I0 is the initial
activity.
The half-life (T) is the time taken for the activity (or the number of
undecayed nuclei) to halve. This is inversely proportional to , as can be
seen:
1
2
 exp T 
ln 12  T
ln 2  T
ln 2
T
.
(12)

If a half-life is too long to measure directly, the value of  can be
determined if I and N are known separately. I would be measured
simply by counting the decays in one year (say), while N would be
measured by putting a fraction of the sample through a mass
spectrometer.
7.4.3
Nuclear Reactions
Now for the final technique: You will need to be able to calculate the
energy released in a nuclear reaction. For this, add up the mass you
started with, and add up the mass at the end. Some mass will have
gone missing. Remembering that mass and energy are basically the
same thing – the ‘lost mass’ is the energy released from the nuclei.
Page 123
Corrections March 2007
2
H
H
4
He
n
3
2.014102
3.016049
4.002604
1.008665
Page 125
Corrections March 2007
8 Practical Physics
8.1 Errors, and how to make them29
Every dog has its day, every silver lining has its cloud, and every
measurement has its error.
If you doubt this, take (sorry – borrow with permission) a school metre
stick, and try and measure the length of a corridor in your school. Try
and measure it to the nearest centimetre. Then measure it again.
Unless you cheated by choosing a short corridor, you should find that
the measurements are different. What’s gone wrong?
Nothing has gone wrong. No measurement is exact, and if you take a
series of readings, you will find that they cluster around the ‘true value’.
This spread of readings is called random error – and will be determined
by the instrument you use and the observation technique. To be more
precise and polite, this kind of ‘error’ is usually called uncertainty, as this
word doesn’t imply any mistake or incompetence on the part of the
scientist.
So, whenever you write down a measurement, you should also write
down its uncertainty. This can be expressed in two ways – absolute and
relative.
The absolute uncertainty gives the size of the spread of readings. You
might conclude that your corridor was (12.3±0.2)m long. In other words,
your measurements are usually within 20cm of 12.3m. In this case the
absolute uncertainty is 20cm.
The absolute uncertainty only gives part of the story. A 10cm error in the
length of a curtain track implies sloppy work. A 10cm error in the total
length of the M1 motorway is an impressive measurement. To make this
clearer, we often state errors (or uncertainties) in percentage form – and
this is called relative uncertainty. The relative uncertainty in the length of
the corridor is
29
A mathematician would probably be appalled at some of the statements I make. The study
of errors and uncertainties is embedded in statistics, which is a well-established discipline.
There are many refinements to the results I quote which are needed to satisfy the rigour of a
professional statistician. However, the thing about uncertainties in measurements is that
quoting them to more than one significant figure is missing the point, and therefore our
methods only need to be accurate to this degree. If you are doing statistics and you want to
take things more seriously, then you will understand (2) from the addition of variances; and
you will realise that in 8.2.1 we really ought to be adding variances not errors. You will also
appreciate that (2) ought to have an (n-1) in the denominator to take into account the
difference between population and sample statistics, and that our section 8.2.2 is a form of
the Binomial theorem to first order.
Page 126
Corrections March 2007
Relative Error 
Abs. Uncertainty 0.2 m

 1.6%  2% .
Measuremen t
12.3 m
(1)
Notice the rounding off at the end. It is usually pointless to give
uncertainties to more than one significant figure.
Every measurement has its uncertainty, and the only way of determining
this is to take more than one measurement, and work out the standard
deviation – to measure the spread. In practice the spread can be
‘eyeballed’ rather than calculated. If the measurements were 54.5cm,
54.7cm and 54.3cm, then there is no need to use a calculator and the
technical definition of deviation. The observation that the spread is
about ±0.2cm is perfectly good enough.
Notice that the more readings you take, the better idea you get of the
spread of the measurements – and hence the better estimate you can
make for the middle, which is indicative of ‘true’ value. Therefore we
find, from statistics, that if you take n measurements, and the absolute
uncertainty is x, then the uncertainty of the mean of those
measurements is approximately: 30
Uncertainty of mean 
x
n
.
(2)
Therefore, the more measurements you take, the more accurate the
work. Notice that if you wish to halve the uncertainty, you need to take
four times as many readings. This is subject to one proviso:
Measurements also have a resolution.
This is the smallest
distinguishable difference that the measuring device (including the
technique) can detect. For a simple length measurement with a metre
ruler, the resolution is probably 1mm. However if, by years of practice
with a magnifying lens, you could divide millimetres into tenths by eye,
you would have a resolution of 0.1mm using the same metre stick. That
is why we say that the resolution depends on the technique as well as on
the apparatus.
The uncertainty of a measurement can never be less than the resolution.
This is the proviso we mentioned below equation (2). Why should this
be the case? Let us have a parable.
Many years ago, the great nation of China had an emperor. The masses
of the population were not permitted to see him. One day, a citizen had
30
This result will be proved in any statistics textbook. To give a brief justification – the more
readings you take, the more likely you are to have some high readings cancelling out some
low readings when you take the average.
Page 127
Corrections March 2007
the sudden desire to know the length of the emperor’s nose. He could
not do this directly, since he was not permitted to visit the emperor. So,
using the apparatus of the imperial administration, he asked all the
regional mandarins to ask the entire population to make a guess. Each
person would make some guess at the imperial nasal length – and the
error of each guess would probably be no more than ±2cm – since nose
lengths tend not to vary by more than about 4cm.
However, the mean would be a different matter. Averaged over the
1000 million measurements, the error in the mean would be 0.7m. So
the emperor’s nose had been measured incredibly accurately – without a
single observation having been made!
The moral of the story: uncertainties are reduced by repeated
measurement, but the error can never be reduced below the resolution
of the technique – here 2cm – since ignorance can not be circumvented
by pooling it with more ignorance.
8.2 Errors, and how to make them worse
Errors are one thing. The trouble is that usually we want to put our
measurements into a formula to calculate something else. For example,
we might want to measure the strength of a magnetic field by measuring
the force on a current-carrying wire B  F IL .
If there is a 7% uncertainty in the current, 2% in the force and 1% in the
length – what is the uncertainty in the magnetic field?
There are two rules you need:
8.2.1
Rule 1 – Adding or subtracting measurements
If two measurements are added or subtracted, the absolute uncertainty
in the result equals the sum (never the difference) of the absolute
uncertainties of the individual measurements.
Therefore if a car is (3.2±0.1)m long, and a caravan is (5.2±0.2)m long,
the total length is (8.4±0.3)m long. Similarly if the height of a two-storey
house is (8.3±0.2)m and the height of the ground floor is (3.1±0.1)m, the
height of the upper floor is (5.2±0.3)m.
Even in the second case, we do not subtract the uncertainties, since
there is nothing stopping one measurements being high, while the other
is low. 31
31
Of course, there is a good chance that the errors will partly cancel out, and so our method
of estimating the overall error is pessimistic. Nevertheless, this kind of error analysis is good
enough for most experiments – after all it is better to overestimate your errors. If you want to
do more careful analysis, then you work on the principle that if the absolute uncertainties in a
Page 128
Corrections March 2007
8.2.2
Rule 2 – Multiplying or dividing measurements
If two measurements are multiplied or divided, the relative uncertainty in
the result equals the sum (never the difference) of the relative
uncertainties of the individual measurements.
Therefore if the speed of a car is 30mph ± 10%, and the time for a
journey is 6 hours ± 2%, the uncertainty in the distance travelled is 12%.
Notice that one consequence of this is that if a measurement, with
relative uncertainty p% is squared (multiplied by itself), the relative error
in the square is 2p% - i.e. doubled. Similarly if the error in measurement
L is p%, the error in Ln is p×n%. Notice that while a square root will
halve the relative error, an inverse square (n=–2) doubles it. All the
minus sign does is to turn overestimates into underestimates. It does
not reduce the magnitude of the relative error. 32
Now we can answer our question about the magnetic field measurement
at the beginning of the section. All three relative errors (in length, force
and current) must be added to give the relative error in the magnetic
field, which is therefore 10%.
8.3 Systematic Errors
All the ‘errors’ mentioned so far are called ‘random’, since we assume
that the measurements will be clustered around the true value. However
often an oversight in our technique will cause a measurement to be
overestimated more often than underestimated or vice-versa. This kind
of error is called ‘systematic error’, and can’t be reduced by averaging
readings. The only way of spotting this kind of error (which is a true
error in that there is something wrong with the measurement) is to repeat
the measurement using a completely different technique, and compare
the results. Just thinking hard about the method can help you spot some
set of measurements are A, B, C…, then the absolute uncertainty in the sum (or in any of the
differences) is given by A  B  C  . This result comes from statistics, where we find
that the variance (the square of the standard deviation) of a sum is equal to the sum of the
variances of the two measurements.
2
32
2
2
The conclusions of this paragraph can be justified using calculus. If measurement x has
absolute uncertainty x, and y (a function of x) is given by y  Ax , then we find that the
relative error in y is given by:
n
y
y

dy x
y x
x
x
 An x n 1
n
n ,
y
x y
x
dx y


that is n multiplied by the relative error in x.
Page 129
Corrections March 2007
systematic errors, but it is still a good idea to perform the experiment a
different way if time allows.
8.4 Which Graph?
You will often have to use graphs to check the functional form of
relationships. You may also have to make measurements using the
graph. In order to do either of these, you usually need to manipulate the
data until you can plot a straight line. A straight line is conclusive proof
that you have got the form of the formula right!
The gradient and y-intercept can then be read, and these enable other
measurements to be made. For example, your aim may be to measure
the acceleration due to gravity. You may plot velocity of falling against
time, in which case you will need to find the gradient of the line.
At its most general, you will have a suspected functional form y=f(x), and
you will need to work out what is going on in the function f. Notice that
our experiment will give us pairs of (x,y) values – what is not known are
the parameters in the function f. We find them by manipulating the
equation:
y  f ( x)

.
g ( x, y )  A h ( x , y )  B
We can then plot g(x,y) against h(x,y), and obtain the parameters A and
B from the gradient and intercept of the line. Furthermore, the presence
of the straight line on the graph assures us that our function f was a
good guess. We shall now look at the most common examples.
8.4.1
Exponential growth or decay
Here we have the functional form y  Ae Bx , where A and B need to be
determined. We manipulate the equation:
y  Ae Bx
.
ln y  ln A  Bx
So we plot (ln y) on the vertical axis, and (x) on the horizontal. The yintercept gives ln A, and the gradient gives B.
8.4.2
Logarithmic growth or decay
Here we have the functional form y  A  B ln x , and again, we need to
work out the values of A and B. This equation is already in linear form –
we plot y on the vertical, and (ln x) on the horizontal. The y-intercept
gives A, and the gradient gives B.
Page 130
Corrections March 2007
8.4.3
Power laws
This covers all equations with unknown powers:
manipulation involves logarithms:
y  Ax B .
The
y  Ax B
ln y  ln A  ln x B .
ln y  ln A  B ln x
Here we plot (ln y) against (ln x), and find the power (B) as the gradient
of the line. The A value can be inferred from the y-intercept, which is
equal to ln A.
8.4.4
Other forms
Even hideous looking equations can be reduced to straight lines if you
crack the whip hard enough. How about y  A x  Bx 3 ? Is it tasty
enough for your breakfast? Actually it’s fine if digested slowly:
y  A x  Bx 3
.
y
 A  Bx 5 2
x
This looks even worse, doesn’t it? But remember that it is x and y that
are known. If we plot ( y x ) on the vertical, and (x5/2) on the horizontal,
a straight line appears, and we can read A and B from the y-intercept
and gradient respectively.
8.5 Questions
1. Work out the relative uncertainty when a 5V battery is measured to the
nearest 0.2V.
2. If I don’t want to have to correct my watch more than once a week, and I
never want my watch to be more than 1s from the correct time, calculate
the necessary maximum relative uncertainty of the electronic oscillator
which I can tolerate.
3. My two-storey house is 7.05±0.02m tall. The ground floor is 3.2±0.01m
tall. How tall is the first floor?
4. I want to measure the resistance of a resistor. My voltmeter can read up
to 5V, with an absolute uncertainty of 0.1V. My ammeter can read up to
1A with an absolute uncertainty of 0.02A. Assuming that my resistor is
approximately 10Ω, calculate the absolute uncertainty of the resistance I
measure using the formula R=V/I. Assume that I choose the current to
make the relative uncertainty as small as possible.
Page 131
Corrections March 2007
9 Appendix
9.1 Multiplying Vectors
Physics is riddled with quantities which have both magnitude and
direction – velocity, acceleration, displacement, force, momentum,
angular velocity, torque, and electric field to name but eight. When
describing these, it is very useful to use vector notation. At best this
saves us writing out separate equations for each of the components.
While the addition and subtraction of vectors is reasonably
straightforward (you add, or subtract, the components to get the
components of the result), multiplication is more tricky.
You can think of a vector as a little arrow. You can add them by stacking
them nose-to-tail, or subtract them by stacking them nose-to-nose. But
how do you go about multiplying them? It is not obvious!
To cut a long story short, you can’t do it unambiguously. However there
are two vector operators which involve multiplication and are useful in
physics. Ordinary multiplication is commonly written with either the
cross (×) or dot (●), so when it comes to vectors we call our two different
‘multiplication’ processes the dot product and cross product to
distinguish them.
These are the closest we get to performing
multiplication with vectors.
9.1.1
The Dot product (or scalar product)
A ton of bricks is lifted a metre, then moved horizontally by 2m. How
much work is done? Work is given by the product of force and distance,
however only the vertical lifting (not the horizontal shuffling) involves
work. In this case the work is equal to the weight (about 9.8kN)
multiplied by the vertical distance (1m).
This gives us one useful way of ‘multiplying’ vectors – namely to multiply
the magnitude of the first, by the component of the second which is
parallel to the first.
If the two vectors are A and B, with magnitudes A and B, and with an
angle  between them, then the component of B parallel to A is Bcos.
Therefore the dot product is given by ABcos.
A  B  AB cos 
(1)
Notice that the dot product of two vectors is itself a scalar. Note that
when we talk about the square of a vector, we mean its scalar product
with itself. Since in this case, =0, this is the same as the square of the
vector’s magnitude.
Page 132
Corrections March 2007
The dot product is also commutative, in other words, the order of the two
vectors A and B does not matter, since A●B=B●A.
The dot product of two vectors written using Cartesian co-ordinates is
particularly easy to calculate. If we use i, j, and k to represent the unit
vectors pointing along the +x, +y and +z axes, then
ai  bj  ck   ui  vj  wk   aui  i  avi  j  awi  k
 buj  i  bvj  j  bwj  k
 cuk  i  cvk  j  cwk  k
 au  bv  cw
9.1.2
(2)
The Cross product (or vector product)
If the dot product produced a scalar, what are we to do if a vector is
needed as the result of our multiplication? Answer: a cross product.
Our first dilemma is to choose the direction of the result. Given that the
vectors will, in general, not be parallel or antiparallel, we can’t choose
the direction of one of them – that would not be fair! The two vectors will
usually define a plane, so perhaps we could use a vector in this plane as
the result? No, that wouldn’t do either – there is still an infinite number
of directions to choose from! A solution is presented if we choose the
vector perpendicular to this plane. This narrows the choice down to two
directions – and we use a convention to choose which.
Notice that the result of the cross product must be zero if the two vectors
are parallel, since in this case we can’t define a plane using the vectors.
It follows that the cross product of a vector with itself is zero. This
means that we aren’t going to be interested in the component of the
second vector which is parallel to the first when calculating the product.
On the contrary, it is the perpendicular component which matters.
The cross product of two vectors is defined as the magnitude of the first,
multiplied by the component of the second which is perpendicular to the
first. The product is directed perpendicular to both vectors. To be more
precise, imagine a screw attached to the first vector. The cross product
goes in the direction the screw advances when the first vector is twisted
to line up with the second. The cross product of a vector lying along the
+x axis with one lying along the +y axis is one lying along the +z axis.
The cross product of ‘up’ with ‘forwards’ is ‘left’.
A  B  AB sin 
(3)
With a definition as obtuse as this, you could be forgiven for wondering
whether it had any practical use at all! However they turn out to be very
useful in physics – especially when dealing with magnetic fields and
rotational motion.
Page 133
Corrections March 2007
Notice that, where the dot product was commutative, the cross product is
anticommutative. In other words, A×B=–B×A, so make sure you don’t
swap the vectors over inadvertently.
The vector product of vectors written in Cartesian form can also be
calculated:
ai  bj  ck   ui  vj  wk   aui  i  avi  j  awi  k
 buj  i  bvj  j  bwj  k
 cuk  i  cvk  j  cwk  k
 0  avk  awj
 buk  0  bwi
(4)
 cuj  cvi  0
 bw  cv i  cu  awj  av  bu k
i
j k
a b c
u v w
where the most convenient way of remembering the result is as the
determinant of the 3×3 matrix shown.
Page 134
Corrections March 2007
9.2 Dimensional Analysis
If you look back at the ‘flow equation’ (24 in section 1.3.3), you will see
something interesting about the units.
Current (A) = Charge density (C/m3)  Area (m2)  Speed (m/s)
If we ‘do algebra’ with the units on the right hand side, we get
C
m C
 m2    A ,
3
m
s s
and this agrees with the units of the left hand side. Now this may all
seem pretty obvious, but it gives us a useful procedure for checking
whether our working is along the right lines. If, during your calculations,
you find yourself adding a charge of 3C to a distance of 6m to get a
result of 9N; or you multiply a speed of 13m/s by a time of 40s and get a
current of 520A; then in either case you must have made a mistake!
We can also use the principle that units must balance to guess the form
of an equation we do not know how to derive. For example, you may
guess that the time period of a simple pendulum might depend on the
length of the pendulum L, the strength of the local gravity g and the
mass of the pendulum bob m. Now




L is measured in m
g is measured in N/kg or m/s2
m is measured in kg,
and we want a time period, which will be measured in s.
The only way these measurements can be combined to make something
in seconds is to take L, divide it by g (this gives something in s2) and
then take the square root. Therefore, without knowing any physics of the
simple harmonic oscillator, we have shown that the time period of a
pendulum is related to L g and will be independent of the mass m.
Similarly, notice what happens if you multiply ohms by farads:
F 
V C C
   s.
A V A
Yes, you get seconds. Therefore, it should come as little surprise to you
that if you double the resistance of a capacitor-resistor network, it will
take twice as long to charge or discharge. Furthermore, you have
worked this out without recourse to calculus or the tedious electrical
details of section 6.1.2.2.
Page 135
Corrections March 2007
Of course, one drawback of the method presented here is that some
quantities have two different units. For example, gravitational field
strength could be N/kg or m/s2. Electric field strength could be N/C or
V/m. How do you know which to choose? The answer is that if you
restrict yourself to using the minimum number of units in your working,
and express all others in terms of them, you will not have any difficulties.
Usually people choose m, s, kg and A but any other combination of
independent units 33 will do equally well. 34
In books you may see folk use L, T, M, I and  to represent the
‘dimensions’ of length, time, mass, electric current and temperature.
This is just a more formal way of doing what we have done here using
the S.I. units. In these books, the dimension of speed would be written
as
[speed] = L T1,
and the dimensions of force would be written
[force] = M L T2,
where the square brackets mean ‘dimensions of’. Technically, this is
more correct than using the S.I. units, because some quantities are
dimensionally the same, but have very different meanings (and hence
units). For example, torque and energy have the same dimensions, but
you wouldn’t want to risk confusing them by using the same unit for both.
Similarly an angle in radians has no dimensions at all (being a an arc
length in metres divided by a radius in metres), but we wouldn’t want to
confuse it with an ordinary number like 3. 35
33
By independent we mean that no one unit can be derived entirely from a combination of the
others. For example, m, s, kg and J would be no good as a set of four since J can already be
expressed in terms of the others J = kg  (m/s)2, and hence we have ambiguity arising as to
how we express quantities.
34
OK, if you want to do work where there are electric currents and temperatures as well as
mechanical quantities, you might need to go up to five (an extra one for temperature).
35
Indeed, quantities with ‘no units’ are usually said to have the dimensions of the number
one. Thus [angle] = 1. It follows that angular velocity has dimensions of [angle][time] = 1T
= T1. This comes from the property 1 has in being the ‘unity’ operator for multiplication.
Page 136